b2ar_salmeterol-analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from detect_intermediates import *\n",
    "from interpret_tICs import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/enf/anaconda3/lib/python3.5/site-packages/matplotlib/__init__.py:1350: UserWarning:  This call to matplotlib.use() has no effect\n",
      "because the backend has already been chosen;\n",
      "matplotlib.use() must be called *before* pylab, matplotlib.pyplot,\n",
      "or matplotlib.backends is imported for the first time.\n",
      "\n",
      "  warnings.warn(_use_error_msg)\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# changing matplotlib the default style\n",
    "matplotlib.style.use('ggplot')\n",
    "\n",
    "from PDB_Order_Fixer import PDB_Order_Fixer\n",
    "import mdtraj as md\n",
    "import os\n",
    "import numpy as np\n",
    "import h5py\n",
    "\n",
    "import datetime\n",
    "import glob\n",
    "import copy\n",
    "from functools import partial \n",
    "import operator\n",
    "import time\n",
    "\n",
    "import random \n",
    "import subprocess\n",
    "from subprocess import Popen\n",
    "import sys\n",
    "from io_functions import *\n",
    "from custom_clusterer import *\n",
    "from custom_tica import *\n",
    "from custom_featurizer import *\n",
    "from pdb_editing import *\n",
    "from analysis import *\n",
    "from io_functions import *\n",
    "#from topology_fixing import *\n",
    "from subsampling import *\n",
    "from conversions import *\n",
    "from custom_msm import *\n",
    "#from grids import *\n",
    "from docking_analysis import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import io_functions\n",
    "from imp import reload\n",
    "reload(io_functions)\n",
    "from io_functions import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from operator import itemgetter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 97, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 265, 266, 267, 268, 269, 270, 271, 273, 274, 275, 276, 277, 278, 279, 280, 281, 282, 283, 284, 285, 286, 287, 288, 289, 290, 291, 292, 293, 294, 295, 296, 297, 298, 299, 300, 301, 302, 303, 304, 305, 306, 307, 308, 309, 310, 311, 312, 313, 314, 315, 316, 317, 318, 319, 320, 321, 322, 323, 324, 325, 326, 327, 328, 329, 330, 331, 332, 333, 334, 335, 336, 337, 338, 339, 340, 341, 342]\n",
      "There are 274 common residues between input structures\n"
     ]
    }
   ],
   "source": [
    "from get_variable_names import *\n",
    "inactive_dir='/home/enf/md_simulations/b2ar/reference_receptors/2RH1_prepped-A.pdb'\n",
    "active_dir='/home/enf/md_simulations/b2ar/reference_receptors/3P0G_pymol_prepped-A.pdb'\n",
    "simulation_structure='/home/enf/md_simulations/b2ar/reference_receptors/ionized.pdb'\n",
    "common_residues_pkl='/home/enf/md_simulations/b2ar/reference_receptors/common_residues2.pkl'\n",
    "contact_residues = find_common_residues([inactive_dir, active_dir, simulation_structure], common_residues_pkl)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import sys\n",
    "import csv\n",
    "\n",
    "\n",
    "# all the parameters\n",
    "\n",
    "exacycle = False\n",
    "\n",
    "n_clusters = 300\n",
    "lag_time = 5\n",
    "if exacycle:\n",
    "  lag_time *= 2\n",
    "\n",
    "msm_lag_time = 5\n",
    "n_components = 10\n",
    "n_samples = 10\n",
    "n_macrostates = 25\n",
    "n_trees = 100\n",
    "\n",
    "precision = \"SP\"\n",
    "\n",
    "sparse = True\n",
    "wolf = True\n",
    "\n",
    "is_sparse = \"\"\n",
    "if(sparse): is_sparse = \"sparse-\"\n",
    "\n",
    "shrinkage_string = \"\"\n",
    "rho_string = \"\"\n",
    "if wolf and not sparse:\n",
    "  wolf_string = \"_wolf_\"\n",
    "  shrinkage = None\n",
    "  shrinkage_string = \"autoShrinkage\"\n",
    "  rho = None\n",
    "  rho_string = \"\"\n",
    "elif wolf and sparse:\n",
    "  wolf_string = \"_wolf_\"\n",
    "  shrinkage = None\n",
    "  shrinkage_string = \"autoShrinkage\"\n",
    "  rho = 0.005\n",
    "  rho_string = \"_rho0pt005\"\n",
    "else:\n",
    "  wolf_string = \"\"\n",
    "  shrinkage = 0.001\n",
    "  shrinkage_string = \"0pt001\"\n",
    "  rho = None\n",
    "  rho_string = \"\"\n",
    "\n",
    "traj_ext = \".h5\"\n",
    "base = \"/home/enf/md_simulations/b2ar/Salmeterol\"\n",
    "traj_dir = \"/home/enf/md_simulations/b2ar/Salmeterol/h5_trajectories\"\n",
    "structure = None\n",
    "iterative = False\n",
    "featurize_parallel = True\n",
    "sampling_method = \"random\"\n",
    "precision = \"SP\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "In this iPython notebook, we will featurize MOR ligand binding simulation by pairwise distances between the ligand and different receptor residues. We will then perform tICA and prospectively build an MSM. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "feature_name=\"contact_6pt6_angstroms\"\n",
    "ori_feature_name = copy.deepcopy(feature_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "lag_time = 5\n",
    "n_components = 10\n",
    "n_clusters = 100\n",
    "\n",
    "rho = 0.01 # tica regularization parameter\n",
    "rho_string = \"0pt01\"\n",
    "lag_time=50"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/enf/md_simulations/b2ar/Salmeterol/featurescontact_6pt6_angstroms-CA-py3-far\n",
      "/home/enf/md_simulations/b2ar/Salmeterol/featurescontact_6pt6_angstroms-CA-py3-far\n"
     ]
    }
   ],
   "source": [
    "schemes = [\"closest-heavy\", \"CA\"]\n",
    "feature_name = \"%s-CA-py3-far\" %ori_feature_name\n",
    "#feature_name = \"%s_phi_psi_chi2\" %feature_name\n",
    "(active_ref_dir, inactive_ref_dir, simulation_ref_dir, scripts_dir,\n",
    "          ligand_dir, agonist_dir, inverse_agonist_dir, biased_agonist_dir, ref_receptors_dir, whole_trajectory_pnas,\n",
    "          sasa_file) = get_base_files(base)\n",
    "\n",
    "tica_dir = get_tica_dir(base, is_sparse, lag_time, n_components, feature_name, \n",
    "                                 wolf_string, shrinkage_string, rho_string)\n",
    "#tica_dir = \"%s-with-inactive\" %tica_dir\n",
    "tica_dir = \"%s-backup\" %tica_dir\n",
    "ori_tica_dir = copy.deepcopy(tica_dir)\n",
    "#tica_dir = \"%s-normalized\" % ori_tica_dir\n",
    "features_dir = get_features_dir(base, feature_name)\n",
    "\n",
    "landmarks_dir = get_landmarks_dir(tica_dir)\n",
    "analysis_dir = get_analysis_dir(tica_dir, n_clusters, sampling_method)\n",
    "gmm_dir = get_gmm_dir(tica_dir)\n",
    "rf_dir = get_rf_dir(tica_dir)\n",
    "\n",
    "\n",
    "ref_tica_dir, ref_tica_coords = get_ref_tica_dirs(tica_dir)\n",
    "\n",
    "graph_file = get_graph_file(tica_dir, msm_lag_time, n_clusters)\n",
    "\n",
    "pnas_titles =  [\"tm6_tm3_dist\", \"rmsd_npxxy_inactive\", \"rmsd_npxxy_active\", \"rmsd_connector_inactive\", \"rmsd_connector_active\"]\n",
    "pnas_features_dir = analysis_dir\n",
    "\n",
    "\n",
    "(clusterer_dir, msm_model_dir, macrostate_dir, features_known, model_dir, projected_features_dir,\n",
    "         projection_operator_dir, ktica_fit_model_filename, ktica_projected_data_filename, nystroem_data_filename,\n",
    "         mutual_information_csv, pearson_csv) = get_tica_files(base, tica_dir, n_clusters, msm_lag_time, n_macrostates)\n",
    "\n",
    "(standardized_features_dir, feature_residues_csv, feature_residues_pkl,\n",
    "          contact_csv, ref_features_dir) = get_feature_files(features_dir)\n",
    "\n",
    "(kmeans_csv, tica_coords_csv, features_csv, active_rmsd_dir, inactive_rmsd_dir, active_pnas_dir, inactive_pnas_joined, active_pnas_joined,\n",
    "        clusters_map_file, ktica_clusters_map_file, analysis_file, combined_file, docking_summary, docking_joined, docking_z_scores_csv,\n",
    "        aggregate_docking, aggregate_docking_joined, docking_pnas_joined, aggregate_docking_pnas, aggregate_docking_pnas_joined, docking_multiple_ligands,\n",
    "        docking_distances_file, docking_pdf, mmgbsa_docking_distances, pnas_coords, mmgbsa_dir, mmgbsa_csv, mmgbsa_pdf, aggregate_mmgbsa,\n",
    "        aggregate_mmgbsa_joined, aggregate_mmgbsa_pnas_joined, mmgbsa_z_scores_csv, active_clusters_csv, intermediate_clusters_csv,\n",
    "        inactive_clusters_csv, pnas_clusters_averages, tica_clusters_averages, tica_classes_csv, tica_samples_csv, subgraph_save_base,\n",
    "        degree_save_base, degree_map_csv, degree_z_map_csv, aggregate_docking_pnas_degree_z_joined, tic_residue_csv, feature_coefs_csv,\n",
    "        duplicated_feature_coefs_csv) = get_analysis_files(analysis_dir, n_clusters, tica_dir, tica_dir, sampling_method, n_samples, precision,\n",
    "                                                           msm_lag_time)\n",
    "\n",
    "(inactive_pnas_distances_dir, active_pnas_distances_dir, active_pnas_all_distances_dir,\n",
    "          inactive_pnas_distances_new_csv, active_pnas_distances_new_csv, active_pnas_joined, active_pnas_means, pnas_coords_dir,\n",
    "          pnas_coords_csv, pnas_all_coords_csv, pnas_coords_hexbin_dir, pnas_coords_co_crystallized_docking_dir,\n",
    "          pnas_coords_active_colors_dir, user_defined_features_file, reaction_coordinates_trajs_file) = get_pnas_files(whole_trajectory_pnas, pnas_features_dir)\n",
    "\n",
    "features_dir = get_features_dir(base, feature_name)\n",
    "\n",
    "\n",
    "\n",
    "graph_file = get_graph_file(tica_dir, msm_lag_time, n_clusters)\n",
    "(scripts_dir, pymol_fixpdb_dir) = get_script_dir(scripts_dir)\n",
    "(save_dir, reimaged_dir, mae_dir, combined_reimaged_dir, grid_dir, docking_dir) = get_docking_dirs(tica_dir, n_clusters, n_components, n_samples, sampling_method, precision)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"\\nfrom ipyparallel import Client\\nrc = Client()\\nprint(len(rc.ids))\\ndview = rc[:]\\ndview.map(os.chdir, ['/home/enf/b2ar_analysis/conformation']*len(rc.ids))\\n\""
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "from ipyparallel import Client\n",
    "rc = Client()\n",
    "print(len(rc.ids))\n",
    "dview = rc[:]\n",
    "dview.map(os.chdir, ['/home/enf/b2ar_analysis/conformation']*len(rc.ids))\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'\\nuser_defined_features_dir = \"%s/user_defined_features\" %traj_dir\\nif not os.path.exists(user_defined_features_dir):\\n    os.makedirs(user_defined_features_dir)\\ncompute_user_defined_features_wrapper(traj_dir, traj_ext, inactive_dir, active_dir, None,\\n                                        feature_name_residues_dict, user_defined_features_file, worker_pool=None, \\n                                        save_dir=user_defined_features_dir, parallel=True, overwrite=False)\\n'"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "user_defined_features_dir = \"%s/user_defined_features\" %traj_dir\n",
    "if not os.path.exists(user_defined_features_dir):\n",
    "    os.makedirs(user_defined_features_dir)\n",
    "compute_user_defined_features_wrapper(traj_dir, traj_ext, inactive_dir, active_dir, None,\n",
    "                                        feature_name_residues_dict, user_defined_features_file, worker_pool=None, \n",
    "                                        save_dir=user_defined_features_dir, parallel=True, overwrite=False)\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'\\nref_user_defined_features_file = \"/home/enf/md_simulations/MOR/h8_reimaged/all_pnas_features/ref_coords.h5\"\\nref_user_defined_features_dir = \"%s/ref_user_defined_features\" %traj_dir\\nif not os.path.exists(ref_user_defined_features_dir):\\n    os.makedirs(ref_user_defined_features_dir)\\n\\ncompute_user_defined_features_wrapper(ref_receptors_dir, \".pdb\", inactive_dir, active_dir, None,\\n                                        feature_name_residues_dict, ref_user_defined_features_file, worker_pool=None, \\n                                        save_dir=ref_user_defined_features_dir, parallel=False, overwrite=True)\\n'"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "ref_user_defined_features_file = \"/home/enf/md_simulations/MOR/h8_reimaged/all_pnas_features/ref_coords.h5\"\n",
    "ref_user_defined_features_dir = \"%s/ref_user_defined_features\" %traj_dir\n",
    "if not os.path.exists(ref_user_defined_features_dir):\n",
    "    os.makedirs(ref_user_defined_features_dir)\n",
    "\n",
    "compute_user_defined_features_wrapper(ref_receptors_dir, \".pdb\", inactive_dir, active_dir, None,\n",
    "                                        feature_name_residues_dict, ref_user_defined_features_file, worker_pool=None, \n",
    "                                        save_dir=ref_user_defined_features_dir, parallel=False, overwrite=True)\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "already featurized\n",
      "structure\n",
      "/home/enf/md_simulations/b2ar/reference_receptors/2RH1_prepped-A.pdb\n",
      "structure\n",
      "None\n",
      "mdraj_index_combinations[0:10]\n",
      "[(29, 56), (29, 166), (29, 235), (29, 117), (29, 271), (29, 17), (29, 8), (29, 52), (29, 176), (29, 182)]\n",
      "contact_features[0:10]\n",
      "[[ALA57, LEU84], [ALA57, PHE194], [ALA57, ASN293], [ALA57, LEU145], [ALA57, SER329], [ALA57, LEU45], [ALA57, MET36], [ALA57, LEU80], [ALA57, SER204], [ALA57, VAL210]]\n",
      "About to compute 36336 features\n",
      "(1, 36336)\n",
      "cutoff\n",
      "0.66\n",
      "distances[0:10]\n",
      "[[ 1.54299212  3.76996589  3.27061844 ...,  2.0320251   1.37413311\n",
      "   2.02950144]]\n",
      "There are 1046 residue-residue contacts below cutoff in structure.\n",
      "structure\n",
      "/home/enf/md_simulations/b2ar/reference_receptors/3P0G_pymol_prepped-A.pdb\n",
      "structure\n",
      "None\n",
      "mdraj_index_combinations[0:10]\n",
      "[(34, 61), (34, 171), (34, 233), (34, 122), (34, 269), (34, 22), (34, 13), (34, 57), (34, 181), (34, 187)]\n",
      "contact_features[0:10]\n",
      "[[ALA57, LEU84], [ALA57, PHE194], [ALA57, ASN293], [ALA57, LEU145], [ALA57, SER329], [ALA57, LEU45], [ALA57, MET36], [ALA57, LEU80], [ALA57, SER204], [ALA57, VAL210]]\n",
      "About to compute 36336 features\n",
      "(1, 36336)\n",
      "cutoff\n",
      "0.66\n",
      "distances[0:10]\n",
      "[[ 1.49192929  3.74536014  3.24382639 ...,  1.97115088  1.39178336\n",
      "   1.95263052]]\n",
      "There are 1024 residue-residue contacts below cutoff in structure.\n",
      "structure\n",
      "/home/enf/md_simulations/b2ar/reference_receptors/ionized.pdb\n",
      "structure\n",
      "None\n",
      "mdraj_index_combinations[0:10]\n",
      "[(3561, 3588), (3561, 3698), (3561, 3769), (3561, 3649), (3561, 3805), (3561, 3549), (3561, 3540), (3561, 3584), (3561, 3708), (3561, 3714)]\n",
      "contact_features[0:10]\n",
      "[[ALA57, LEU84], [ALA57, PHE194], [ALA57, ASN293], [ALA57, LEU145], [ALA57, SER329], [ALA57, LEU45], [ALA57, MET36], [ALA57, LEU80], [ALA57, SER204], [ALA57, VAL210]]\n",
      "About to compute 36336 features\n",
      "(1, 36336)\n",
      "cutoff\n",
      "0.66\n",
      "distances[0:10]\n",
      "[[ 1.46443534  3.66438365  3.21374774 ...,  1.90092862  1.36685383\n",
      "   2.05382776]]\n",
      "There are 1047 residue-residue contacts below cutoff in structure.\n",
      "dihedral_residues:\n",
      "[]\n"
     ]
    }
   ],
   "source": [
    "import custom_featurizer\n",
    "reload(custom_featurizer)\n",
    "from custom_featurizer import *\n",
    "cutoff = 0.66\n",
    "\n",
    "# output\n",
    "# ---> permutations of tuple pairs (res1, res2)\n",
    "# |   distance(res1, res2) at time t\n",
    "# |\n",
    "# time t\n",
    "featurize_contacts_custom(traj_dir, features_dir = features_dir, traj_ext = traj_ext, contact_residue_pairs_file = feature_residues_pkl, structures=[inactive_dir, active_dir, simulation_structure], contact_residues=contact_residues,\n",
    "                          residues_map = None, contact_cutoff = cutoff, parallel=True, exacycle = exacycle, traj_top_structure = None, iterative=False,\n",
    "                          user_specified_atom_residue_pairs = [], load_from_file=False, worker_pool=None, schemes=schemes, binarize=None, redo=False, within_turn=False) # removed worker_pool = dview, added parallel = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# returns ordered list of every dataset file (features)\n",
    "get_trajectory_files(features_dir, \".dataset\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# featurized_trajs = load_file_list(get_trajectory_files(features_dir, \".dataset\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# [(featurized_trajs[i].shape, get_trajectory_files(traj_dir, traj_ext)[i]) for i in range(0, len(featurized_trajs))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "n = StandardScaler()\n",
    "for feature in featurized_trajs:\n",
    "    n.partial_fit(feature)\n",
    "# normalizing features\n",
    "# transforms features to zero mean and unit variance (z-scoring)\n",
    "save_file = \"%s/normalizer.h5\" %features_dir\n",
    "verbosedump(n, save_file)\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import pickle\n",
    "with open(feature_residues_pkl, \"rb\") as f:\n",
    "    feature_residues = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "dir(feature_residues[0])\n",
    "dir(feature_residues[0].residue_i)\n",
    "feature_residues[0].residue_i.resSeq"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "import custom_tica\n",
    "reload(custom_tica)\n",
    "from custom_tica import *\n",
    "# fit the tica model, apply model to each matrix\n",
    "fit_and_transform(features_directory = features_dir, model_dir = tica_dir, stride=None, lag_time = lag_time, \n",
    "                  n_components = n_components, sparse = sparse, wolf = wolf, rho = rho, shrinkage = shrinkage,\n",
    "                  parallel=False, traj_ext = traj_ext, normalize=True,\n",
    "                  partial_fit=False, recompute_tica = False) # deleted features=featurized_trajs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#bw_trajectories = [t for t in get_trajectory_files(traj_dir, \".h5\") if \"RUN\" in t] # fixing \"chain\" problem"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# chain swapping\n",
    "\n",
    "\"\"\"\n",
    "from functools import partial\n",
    "from importlib import reload\n",
    "import custom_featurizer\n",
    "reload(custom_featurizer)\n",
    "from custom_featurizer import swap_chains_in_traj\n",
    "\n",
    "swap_chains_in_traj_partial = partial(swap_chains_in_traj, chains=(\"I\", \"A\"))\n",
    "dview.map_sync(swap_chains_in_traj_partial, bw_trajectories)\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# [(c.id, len([r for r in c.residues if r.is_protein])) for c in t.topology.chains]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# featurize reference receptors (controls)\n",
    "\n",
    "import custom_featurizer\n",
    "reload(custom_featurizer)\n",
    "from custom_featurizer import *\n",
    "\n",
    "ref_receptors_dir = '/home/enf/md_simulations/b2ar/reference_receptors'\n",
    "featurize_contacts_custom(ref_receptors_dir, features_dir = ref_features_dir, traj_ext = \".pdb\", contact_residue_pairs_file = feature_residues_pkl, structures=[inactive_dir, active_dir, simulation_structure], contact_residues=contact_residues,\n",
    "                          residues_map = None, contact_cutoff = cutoff, parallel=False, exacycle = exacycle, traj_top_structure = None, iterative=False,\n",
    "                          user_specified_atom_residue_pairs = [], load_from_file=False, worker_pool=None, schemes=schemes, binarize=None, redo=False, within_turn=False)\n",
    "\n",
    "#featurize_contacts_custom(ref_receptors_dir, features_dir = ref_features_dir, traj_ext = \".pdb\", contact_residue_pairs_file = feature_residues_pkl, structures=[inactive_dir, active_dir], contact_residues=contact_residues,\n",
    "#                          residues_map = None, contact_cutoff = cutoff, parallel = False, exacycle = exacycle, traj_top_structure = None, iterative=False,\n",
    "#                          user_specified_atom_residue_pairs = [], load_from_file=False, worker_pool=None, schemes=schemes, within_turn=False, binarize=None, redo=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "ref_features = load_file_list(get_trajectory_files(ref_features_dir, '.dataset'))\n",
    "ref_features\n",
    "for x in ref_features :\n",
    "    print(x.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "transform(existing_model = projection_operator_dir, features_directory = ref_features_dir, tica_dir = ref_tica_dir, normalizer=compat_verboseload(\"%s/normalizer.h5\" %features_dir))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[ALA119 to ALA78,\n",
       " ALA119 to ASP79,\n",
       " ALA119 to LEU124,\n",
       " ALA119 to LEU75,\n",
       " ALA119 to MET82,\n",
       " ALA119 to SER161,\n",
       " ALA119 to SER207,\n",
       " ALA119 to SER74,\n",
       " ALA119 to TRP158,\n",
       " ALA119 to VAL114,\n",
       " ALA119 to VAL157,\n",
       " ALA128 to GLU122,\n",
       " ALA128 to ILE135,\n",
       " ALA128 to ILE214,\n",
       " ALA128 to ILE278,\n",
       " ALA128 to MET215,\n",
       " ALA128 to MET279,\n",
       " ALA128 to PHE133,\n",
       " ALA128 to PRO211,\n",
       " ALA128 to THR123,\n",
       " ALA128 to TYR219,\n",
       " ALA128 to TYR326,\n",
       " ALA128 to VAL216,\n",
       " ALA128 to VAL218,\n",
       " ALA128 to VAL222,\n",
       " ALA134 to GLN142,\n",
       " ALA134 to LEU144,\n",
       " ALA134 to LEU145,\n",
       " ALA134 to LYS140,\n",
       " ALA134 to LYS267,\n",
       " ALA134 to PHE139,\n",
       " ALA134 to SER143,\n",
       " ALA134 to THR68,\n",
       " ALA134 to TYR141,\n",
       " ALA134 to VAL129,\n",
       " ALA150 to LEU145,\n",
       " ALA150 to LEU155,\n",
       " ALA150 to PHE71,\n",
       " ALA150 to SER143,\n",
       " ALA150 to TYR70,\n",
       " ALA150 to VAL126,\n",
       " ALA150 to VAL67,\n",
       " ALA176 to ALA181,\n",
       " ALA176 to ASN196,\n",
       " ALA176 to GLN197,\n",
       " ALA176 to ILE182,\n",
       " ALA176 to PHE194,\n",
       " ALA176 to THR195,\n",
       " ALA176 to TYR185,\n",
       " ALA181 to ALA186,\n",
       " ALA181 to ARG175,\n",
       " ALA181 to ASP192,\n",
       " ALA181 to CYS190,\n",
       " ALA181 to GLU187,\n",
       " ALA181 to PHE194,\n",
       " ALA181 to THR195,\n",
       " ALA186 to ARG175,\n",
       " ALA186 to ASN103,\n",
       " ALA198 to SER203,\n",
       " ALA198 to SER204,\n",
       " ALA198 to TRP173,\n",
       " ALA198 to VAL297,\n",
       " ALA200 to ASN293,\n",
       " ALA200 to HIS296,\n",
       " ALA200 to ILE205,\n",
       " ALA200 to ILE294,\n",
       " ALA200 to PHE193,\n",
       " ALA200 to PHE290,\n",
       " ALA200 to THR195,\n",
       " ALA200 to TYR174,\n",
       " ALA200 to VAL297,\n",
       " ALA202 to GLN197,\n",
       " ALA202 to ILE294,\n",
       " ALA202 to SER207,\n",
       " ALA202 to THR164,\n",
       " ALA226 to ALA271,\n",
       " ALA226 to ARG221,\n",
       " ALA226 to CYS265,\n",
       " ALA226 to GLU268,\n",
       " ALA226 to HIS269,\n",
       " ALA226 to ILE135,\n",
       " ALA226 to LEU266,\n",
       " ALA226 to LEU275,\n",
       " ALA226 to LYS267,\n",
       " ALA226 to TYR141,\n",
       " ALA271 to ARG328,\n",
       " ALA271 to GLY276,\n",
       " ALA271 to LEU266,\n",
       " ALA271 to TYR219,\n",
       " ALA271 to VAL222,\n",
       " ALA335 to ALA57,\n",
       " ALA335 to ARG63,\n",
       " ALA335 to ILE58,\n",
       " ALA335 to LEU340,\n",
       " ALA335 to LEU53,\n",
       " ALA335 to LEU64,\n",
       " ALA335 to PHE61,\n",
       " ALA335 to PRO330,\n",
       " ALA46 to ASN51,\n",
       " ALA46 to GLY320,\n",
       " ALA46 to PHE321,\n",
       " ALA46 to PRO323,\n",
       " ALA46 to SER319,\n",
       " ALA46 to SER41,\n",
       " ALA57 to ASP331,\n",
       " ALA57 to GLN65,\n",
       " ALA57 to GLU62,\n",
       " ALA57 to LEU339,\n",
       " ALA57 to LEU64,\n",
       " ALA57 to PHE332,\n",
       " ALA57 to PHE336,\n",
       " ALA57 to VAL52,\n",
       " ALA59 to GLN65,\n",
       " ALA59 to LEU64,\n",
       " ALA59 to THR73,\n",
       " ALA59 to VAL54,\n",
       " ALA76 to ASN322,\n",
       " ALA76 to ASN51,\n",
       " ALA76 to CYS327,\n",
       " ALA76 to GLY50,\n",
       " ALA76 to ILE47,\n",
       " ALA76 to ILE55,\n",
       " ALA76 to ILE58,\n",
       " ALA76 to PHE71,\n",
       " ALA76 to PRO323,\n",
       " ALA76 to THR123,\n",
       " ALA76 to TYR326,\n",
       " ALA76 to VAL52,\n",
       " ALA76 to VAL54,\n",
       " ALA76 to VAL81,\n",
       " ALA78 to ASN51,\n",
       " ALA78 to CYS116,\n",
       " ALA78 to GLY83,\n",
       " ALA78 to ILE112,\n",
       " ALA78 to ILE121,\n",
       " ALA78 to ILE47,\n",
       " ALA78 to ILE55,\n",
       " ALA78 to LEU115,\n",
       " ALA78 to SER120,\n",
       " ALA78 to SER319,\n",
       " ALA78 to THR118,\n",
       " ALA78 to THR123,\n",
       " ALA78 to THR73,\n",
       " ALA78 to TRP158,\n",
       " ALA78 to VAL117,\n",
       " ALA78 to VAL86,\n",
       " ALA85 to CYS116,\n",
       " ALA85 to GLY90,\n",
       " ALA85 to ILE112,\n",
       " ALA85 to LEU80,\n",
       " ALA85 to TRP105,\n",
       " ALA85 to TRP109,\n",
       " ALA91 to GLY37,\n",
       " ALA91 to LYS97,\n",
       " ALA91 to MET36,\n",
       " ALA91 to MET40,\n",
       " ALA91 to SER41,\n",
       " ALA91 to TRP109,\n",
       " ALA91 to TRP313,\n",
       " ALA91 to TYR316,\n",
       " ALA91 to VAL44,\n",
       " ALA91 to VAL86,\n",
       " ALA92 to LYS97,\n",
       " ALA92 to MET40,\n",
       " ALA92 to PHE101,\n",
       " ALA92 to THR100,\n",
       " ALA92 to TRP313,\n",
       " ALA92 to TRP99,\n",
       " ALA92 to VAL87,\n",
       " ARG131 to GLN142,\n",
       " ARG131 to ILE278,\n",
       " ARG131 to LEU144,\n",
       " ARG131 to LEU145,\n",
       " ARG131 to LEU275,\n",
       " ARG131 to MET279,\n",
       " ARG131 to PHE71,\n",
       " ARG131 to SER143,\n",
       " ARG131 to THR136,\n",
       " ARG131 to THR68,\n",
       " ARG131 to TYR141,\n",
       " ARG131 to TYR219,\n",
       " ARG131 to VAL126,\n",
       " ARG131 to VAL218,\n",
       " ARG131 to VAL222,\n",
       " ARG131 to VAL67,\n",
       " ARG151 to LEU145,\n",
       " ARG151 to MET156,\n",
       " ARG151 to PHE71,\n",
       " ARG151 to THR146,\n",
       " ARG151 to TYR70,\n",
       " ARG175 to ASN103,\n",
       " ARG175 to ASN196,\n",
       " ARG175 to GLN170,\n",
       " ARG175 to GLN197,\n",
       " ARG175 to GLU107,\n",
       " ARG175 to ILE169,\n",
       " ARG175 to ILE182,\n",
       " ARG175 to PHE194,\n",
       " ARG175 to PRO168,\n",
       " ARG175 to THR195,\n",
       " ARG175 to TYR185,\n",
       " ARG175 to TYR199,\n",
       " ARG221 to ILE135,\n",
       " ARG221 to PHE133,\n",
       " ARG221 to THR136,\n",
       " ARG221 to TYR132,\n",
       " ARG221 to VAL216,\n",
       " ARG304 to HIS296,\n",
       " ARG304 to ILE309,\n",
       " ARG304 to LEU310,\n",
       " ARG304 to VAL292,\n",
       " ARG328 to ARG333,\n",
       " ARG328 to HIS269,\n",
       " ARG328 to ILE277,\n",
       " ARG328 to LEU275,\n",
       " ARG328 to LYS270,\n",
       " ARG328 to LYS273,\n",
       " ARG328 to PHE336,\n",
       " ARG328 to PRO323,\n",
       " ARG328 to THR274,\n",
       " ARG333 to ARG63,\n",
       " ARG333 to CYS327,\n",
       " ARG333 to GLU338,\n",
       " ARG333 to LEU64,\n",
       " ARG333 to TYR326,\n",
       " ARG63 to ASN69,\n",
       " ARG63 to ASP331,\n",
       " ARG63 to GLU338,\n",
       " ARG63 to ILE334,\n",
       " ARG63 to ILE58,\n",
       " ARG63 to PHE332,\n",
       " ARG63 to PHE336,\n",
       " ARG63 to PRO330,\n",
       " ARG63 to SER329,\n",
       " ASN103 to ASN183,\n",
       " ASN103 to CYS184,\n",
       " ASN103 to CYS190,\n",
       " ASN103 to CYS191,\n",
       " ASN103 to GLU187,\n",
       " ASN103 to GLU188,\n",
       " ASN103 to PHE108,\n",
       " ASN103 to THR189,\n",
       " ASN103 to TYR185,\n",
       " ASN148 to ILE153,\n",
       " ASN148 to TYR70,\n",
       " ASN183 to CYS190,\n",
       " ASN183 to GLU188,\n",
       " ASN183 to HIS178,\n",
       " ASN196 to HIS172,\n",
       " ASN196 to HIS178,\n",
       " ASN196 to ILE201,\n",
       " ASN196 to THR177,\n",
       " ASN196 to TRP173,\n",
       " ASN196 to TYR174,\n",
       " ASN196 to TYR185,\n",
       " ASN196 to VAL297,\n",
       " ASN293 to ILE201,\n",
       " ASN293 to ILE298,\n",
       " ASN293 to ILE303,\n",
       " ASN293 to PHE193,\n",
       " ASN293 to PRO288,\n",
       " ASN293 to SER203,\n",
       " ASN293 to SER204,\n",
       " ASN293 to THR195,\n",
       " ASN293 to TYR209,\n",
       " ASN293 to TYR308,\n",
       " ASN301 to GLU180,\n",
       " ASN301 to HIS296,\n",
       " ASN301 to VAL295,\n",
       " ASN312 to ASP113,\n",
       " ASN312 to CYS285,\n",
       " ASN312 to MET82,\n",
       " ASN312 to PHE289,\n",
       " ASN312 to PHE290,\n",
       " ASN312 to TRP286,\n",
       " ASN312 to VAL117,\n",
       " ASN312 to VAL292,\n",
       " ASN312 to VAL307,\n",
       " ASN312 to VAL317,\n",
       " ASN318 to ASP79,\n",
       " ASN318 to CYS285,\n",
       " ASN318 to ILE121,\n",
       " ASN318 to ILE278,\n",
       " ASN318 to ILE325,\n",
       " ASN318 to ILE47,\n",
       " ASN318 to LEU124,\n",
       " ASN318 to LEU75,\n",
       " ASN318 to PHE282,\n",
       " ASN318 to PRO323,\n",
       " ASN318 to SER120,\n",
       " ASN318 to THR281,\n",
       " ASN318 to THR283,\n",
       " ASN318 to TRP286,\n",
       " ASN318 to TRP313,\n",
       " ASN322 to ASN51,\n",
       " ASN322 to ASP79,\n",
       " ASN322 to CYS327,\n",
       " ASN322 to ILE121,\n",
       " ASN322 to ILE277,\n",
       " ASN322 to ILE278,\n",
       " ASN322 to ILE72,\n",
       " ASN322 to LEU124,\n",
       " ASN322 to LEU75,\n",
       " ASN322 to MET279,\n",
       " ASN322 to PHE282,\n",
       " ASN322 to PHE336,\n",
       " ASN322 to SER120,\n",
       " ASN322 to THR274,\n",
       " ASN322 to THR281,\n",
       " ASN322 to TRP286,\n",
       " ASN322 to VAL317,\n",
       " ASN322 to VAL54,\n",
       " ASN51 to ASP79,\n",
       " ASN51 to CYS77,\n",
       " ASN51 to GLY320,\n",
       " ASN51 to GLY83,\n",
       " ASN51 to ILE72,\n",
       " ASN51 to LEU75,\n",
       " ASN51 to LEU80,\n",
       " ASN51 to PHE321,\n",
       " ASN51 to PRO323,\n",
       " ASN51 to SER319,\n",
       " ASN51 to THR56,\n",
       " ASN51 to TYR326,\n",
       " ASN51 to VAL81,\n",
       " ASN69 to ASP130,\n",
       " ASN69 to CYS327,\n",
       " ASN69 to ILE58,\n",
       " ASN69 to LEU64,\n",
       " ASN69 to LYS147,\n",
       " ASN69 to PHE332,\n",
       " ASN69 to PHE61,\n",
       " ASN69 to SER329,\n",
       " ASN69 to SER74,\n",
       " ASN69 to TYR141,\n",
       " ASN69 to TYR326,\n",
       " ASN69 to VAL54,\n",
       " ASP113 to MET82,\n",
       " ASP113 to PHE108,\n",
       " ASP113 to PHE166,\n",
       " ASP113 to SER165,\n",
       " ASP113 to THR118,\n",
       " ASP113 to TRP286,\n",
       " ASP113 to TYR316,\n",
       " ASP113 to VAL86,\n",
       " ASP130 to CYS125,\n",
       " ASP130 to GLN142,\n",
       " ASP130 to ILE135,\n",
       " ASP130 to LEU144,\n",
       " ASP130 to LEU145,\n",
       " ASP130 to PHE71,\n",
       " ASP130 to SER143,\n",
       " ASP130 to THR68,\n",
       " ASP130 to TYR141,\n",
       " ASP130 to VAL218,\n",
       " ASP130 to VAL222,\n",
       " ASP130 to VAL67,\n",
       " ASP192 to CYS106,\n",
       " ASP192 to CYS184,\n",
       " ASP192 to GLU180,\n",
       " ASP192 to HIS93,\n",
       " ASP192 to LYS305,\n",
       " ASP192 to THR110,\n",
       " ASP192 to TRP99,\n",
       " ASP192 to TYR174,\n",
       " ASP192 to TYR185,\n",
       " ASP29 to LYS97,\n",
       " ASP29 to VAL34,\n",
       " ASP300 to GLN197,\n",
       " ASP300 to HIS178,\n",
       " ASP300 to VAL292,\n",
       " ASP300 to VAL295,\n",
       " ASP331 to LEU64,\n",
       " ASP331 to PHE336,\n",
       " ASP331 to PHE61,\n",
       " ASP331 to TYR326,\n",
       " ASP79 to CYS116,\n",
       " ASP79 to GLY320,\n",
       " ASP79 to ILE121,\n",
       " ASP79 to ILE47,\n",
       " ASP79 to ILE72,\n",
       " ASP79 to LEU124,\n",
       " ASP79 to LEU84,\n",
       " ASP79 to PHE282,\n",
       " ASP79 to PHE321,\n",
       " ASP79 to PRO323,\n",
       " ASP79 to SER120,\n",
       " ASP79 to SER319,\n",
       " ASP79 to SER74,\n",
       " ASP79 to TRP158,\n",
       " ASP79 to TYR316,\n",
       " ASP79 to TYR326,\n",
       " ASP79 to VAL48,\n",
       " ASP79 to VAL87,\n",
       " CYS106 to CYS184,\n",
       " CYS106 to CYS190,\n",
       " CYS106 to CYS191,\n",
       " CYS106 to GLU188,\n",
       " CYS106 to ILE169,\n",
       " CYS106 to PHE101,\n",
       " CYS106 to PHE193,\n",
       " CYS106 to SER111,\n",
       " CYS106 to TRP99,\n",
       " CYS106 to TYR174,\n",
       " CYS116 to CYS77,\n",
       " CYS116 to ILE121,\n",
       " CYS116 to LEU75,\n",
       " CYS116 to LEU80,\n",
       " CYS116 to MET82,\n",
       " CYS116 to PHE282,\n",
       " CYS116 to SER111,\n",
       " CYS116 to SER161,\n",
       " CYS116 to SER165,\n",
       " CYS116 to SER319,\n",
       " CYS116 to TRP158,\n",
       " CYS116 to VAL81,\n",
       " CYS116 to VAL86,\n",
       " CYS125 to ILE214,\n",
       " CYS125 to LEU212,\n",
       " CYS125 to MET215,\n",
       " CYS125 to MET279,\n",
       " CYS125 to PRO211,\n",
       " CYS125 to SER120,\n",
       " CYS125 to TYR219,\n",
       " CYS125 to TYR326,\n",
       " CYS125 to VAL210,\n",
       " CYS125 to VAL213,\n",
       " CYS125 to VAL216,\n",
       " CYS125 to VAL218,\n",
       " CYS184 to CYS190,\n",
       " CYS184 to CYS191,\n",
       " CYS184 to GLN179,\n",
       " CYS184 to PHE193,\n",
       " CYS184 to PHE194,\n",
       " CYS184 to THR189,\n",
       " CYS190 to GLU180,\n",
       " CYS190 to GLY102,\n",
       " CYS190 to HIS93,\n",
       " CYS190 to PHE101,\n",
       " CYS190 to TRP99,\n",
       " CYS190 to TYR185,\n",
       " CYS191 to GLU107,\n",
       " CYS191 to GLY102,\n",
       " CYS191 to HIS93,\n",
       " CYS191 to LYS305,\n",
       " CYS191 to THR110,\n",
       " CYS191 to TRP105,\n",
       " CYS191 to TRP109,\n",
       " CYS191 to TRP99,\n",
       " CYS265 to LYS227,\n",
       " CYS265 to LYS270,\n",
       " CYS285 to GLY280,\n",
       " CYS285 to GLY315,\n",
       " CYS285 to ILE314,\n",
       " CYS285 to LEU311,\n",
       " CYS285 to MET279,\n",
       " CYS285 to PHE208,\n",
       " CYS285 to PHE290,\n",
       " CYS285 to PHE321,\n",
       " CYS285 to SER319,\n",
       " CYS285 to TRP313,\n",
       " CYS285 to TYR316,\n",
       " CYS285 to VAL317,\n",
       " CYS327 to GLY50,\n",
       " CYS327 to ILE58,\n",
       " CYS327 to ILE72,\n",
       " CYS327 to PHE332,\n",
       " CYS327 to PHE336,\n",
       " CYS327 to THR274,\n",
       " CYS327 to VAL54,\n",
       " CYS341 to PHE336,\n",
       " CYS77 to ILE55,\n",
       " CYS77 to ILE72,\n",
       " CYS77 to MET82,\n",
       " CYS77 to TRP158,\n",
       " GLN142 to ILE135,\n",
       " GLN142 to PHE133,\n",
       " GLN142 to SER137,\n",
       " GLN170 to GLU107,\n",
       " GLN170 to ILE112,\n",
       " GLN170 to PHE104,\n",
       " GLN170 to PHE108,\n",
       " GLN170 to SER111,\n",
       " GLN170 to SER165,\n",
       " GLN170 to TRP109,\n",
       " GLN197 to GLN299,\n",
       " GLN197 to HIS296,\n",
       " GLN197 to ILE294,\n",
       " GLN197 to ILE298,\n",
       " GLN197 to TYR174,\n",
       " GLN197 to VAL297,\n",
       " GLN224 to ILE135,\n",
       " GLN224 to PHE217,\n",
       " GLN224 to TYR219,\n",
       " GLN299 to ILE294,\n",
       " GLN299 to VAL292,\n",
       " GLN337 to LEU342,\n",
       " GLN337 to LEU53,\n",
       " GLN337 to PHE332,\n",
       " GLN337 to PRO330,\n",
       " GLN65 to ILE58,\n",
       " GLN65 to LYS147,\n",
       " GLN65 to LYS60,\n",
       " GLN65 to TYR70,\n",
       " GLU107 to GLY102,\n",
       " GLU107 to HIS172,\n",
       " GLU107 to ILE112,\n",
       " GLU107 to ILE169,\n",
       " GLU107 to MET171,\n",
       " GLU107 to PHE166,\n",
       " GLU107 to SER165,\n",
       " GLU107 to TRP99,\n",
       " GLU107 to TYR174,\n",
       " GLU107 to TYR185,\n",
       " GLU122 to GLY162,\n",
       " GLU122 to ILE127,\n",
       " GLU122 to ILE159,\n",
       " GLU122 to MET156,\n",
       " GLU122 to MET215,\n",
       " GLU122 to PHE208,\n",
       " GLU122 to PHE282,\n",
       " GLU122 to PRO211,\n",
       " GLU122 to SER161,\n",
       " GLU122 to SER207,\n",
       " GLU122 to THR164,\n",
       " GLU122 to TRP158,\n",
       " GLU122 to VAL117,\n",
       " GLU122 to VAL157,\n",
       " GLU122 to VAL160,\n",
       " GLU122 to VAL206,\n",
       " GLU122 to VAL210,\n",
       " GLU180 to LYS305,\n",
       " GLU180 to PHE194,\n",
       " GLU180 to TYR185,\n",
       " GLU187 to GLY102,\n",
       " GLU187 to ILE182,\n",
       " GLU187 to TRP99,\n",
       " GLU188 to GLY102,\n",
       " GLU188 to PHE101,\n",
       " GLU188 to PHE104,\n",
       " GLU188 to THR100,\n",
       " GLU188 to TRP99,\n",
       " GLU225 to GLU268,\n",
       " GLU225 to HIS269,\n",
       " GLU225 to ILE135,\n",
       " GLU225 to LYS140,\n",
       " GLU225 to SER137,\n",
       " GLU225 to SER220,\n",
       " GLU225 to THR136,\n",
       " GLU225 to TYR132,\n",
       " GLU225 to TYR141,\n",
       " GLU268 to LYS227,\n",
       " GLU268 to LYS273,\n",
       " GLU268 to TYR141,\n",
       " GLU30 to GLY35,\n",
       " GLU30 to MET36,\n",
       " GLU306 to LEU311,\n",
       " GLU338 to LYS60,\n",
       " GLU338 to PHE61,\n",
       " GLU62 to LYS147,\n",
       " GLY102 to THR189,\n",
       " GLY162 to LEU115,\n",
       " GLY162 to LEU167,\n",
       " GLY162 to PRO168,\n",
       " GLY162 to SER111,\n",
       " GLY162 to THR118,\n",
       " GLY162 to VAL157,\n",
       " GLY276 to PHE223,\n",
       " GLY276 to THR281,\n",
       " GLY276 to TYR219,\n",
       " GLY280 to LEU124,\n",
       " GLY280 to LEU275,\n",
       " GLY280 to MET215,\n",
       " GLY280 to THR274,\n",
       " GLY315 to GLY320,\n",
       " GLY315 to ILE121,\n",
       " GLY315 to LEU284,\n",
       " GLY315 to LEU310,\n",
       " GLY315 to MET82,\n",
       " GLY315 to PHE282,\n",
       " GLY315 to PHE289,\n",
       " GLY315 to SER120,\n",
       " GLY315 to TRP286,\n",
       " GLY315 to VAL117,\n",
       " GLY320 to GLY50,\n",
       " GLY320 to ILE325,\n",
       " GLY320 to ILE43,\n",
       " GLY320 to ILE47,\n",
       " GLY320 to VAL48,\n",
       " GLY35 to MET40,\n",
       " GLY37 to LEU42,\n",
       " GLY37 to LEU95,\n",
       " GLY37 to TRP32,\n",
       " GLY37 to VAL31,\n",
       " GLY50 to ILE55,\n",
       " GLY50 to LEU324,\n",
       " GLY50 to LEU45,\n",
       " GLY50 to PHE336,\n",
       " GLY50 to PRO323,\n",
       " GLY50 to SER319,\n",
       " GLY50 to TYR326,\n",
       " GLY83 to ILE43,\n",
       " GLY83 to ILE47,\n",
       " GLY83 to LEU45,\n",
       " GLY83 to PRO88,\n",
       " GLY83 to SER319,\n",
       " GLY83 to TYR316,\n",
       " GLY83 to VAL44,\n",
       " GLY83 to VAL48,\n",
       " GLY90 to LEU95,\n",
       " GLY90 to MET40,\n",
       " GLY90 to PHE101,\n",
       " GLY90 to TRP109,\n",
       " GLY90 to TRP313,\n",
       " GLY90 to TRP99,\n",
       " GLY90 to TYR316,\n",
       " HIS172 to LEU167,\n",
       " HIS172 to PHE166,\n",
       " HIS178 to PHE194,\n",
       " HIS178 to THR195,\n",
       " HIS178 to TYR185,\n",
       " HIS269 to ILE135,\n",
       " HIS269 to LYS227,\n",
       " HIS269 to PHE223,\n",
       " HIS269 to THR274,\n",
       " HIS269 to TYR141,\n",
       " HIS269 to TYR219,\n",
       " HIS296 to ILE291,\n",
       " HIS296 to ILE303,\n",
       " HIS296 to LEU302,\n",
       " HIS296 to LYS305,\n",
       " HIS296 to PHE193,\n",
       " HIS296 to PHE194,\n",
       " HIS296 to THR195,\n",
       " HIS296 to TYR308,\n",
       " HIS93 to ILE309,\n",
       " HIS93 to LYS305,\n",
       " HIS93 to PHE101,\n",
       " HIS93 to PRO88,\n",
       " HIS93 to THR100,\n",
       " HIS93 to THR189,\n",
       " HIS93 to TRP109,\n",
       " HIS93 to TRP99,\n",
       " ILE112 to ILE169,\n",
       " ILE112 to MET82,\n",
       " ILE112 to PHE166,\n",
       " ILE112 to SER165,\n",
       " ILE112 to TRP105,\n",
       " ILE112 to VAL117,\n",
       " ILE112 to VAL81,\n",
       " ILE112 to VAL86,\n",
       " ILE121 to LEU212,\n",
       " ILE121 to LEU75,\n",
       " ILE121 to MET215,\n",
       " ILE121 to MET279,\n",
       " ILE121 to PHE208,\n",
       " ILE121 to PHE282,\n",
       " ILE121 to PHE290,\n",
       " ILE121 to PRO211,\n",
       " ILE121 to SER207,\n",
       " ILE121 to THR283,\n",
       " ILE121 to TRP286,\n",
       " ILE121 to TYR209,\n",
       " ILE121 to VAL126,\n",
       " ILE121 to VAL206,\n",
       " ILE121 to VAL210,\n",
       " ILE127 to ILE278,\n",
       " ILE127 to ILE72,\n",
       " ILE127 to LEU275,\n",
       " ILE127 to LEU75,\n",
       " ILE127 to MET279,\n",
       " ILE127 to PHE71,\n",
       " ILE127 to THR68,\n",
       " ILE127 to THR73,\n",
       " ILE127 to TYR132,\n",
       " ILE127 to TYR219,\n",
       " ILE127 to TYR326,\n",
       " ILE127 to TYR70,\n",
       " ILE127 to VAL218,\n",
       " ILE127 to VAL222,\n",
       " ILE135 to LYS140,\n",
       " ILE135 to LYS227,\n",
       " ILE135 to LYS267,\n",
       " ILE135 to PHE223,\n",
       " ILE135 to TYR141,\n",
       " ILE135 to TYR219,\n",
       " ILE135 to VAL222,\n",
       " ILE153 to LEU145,\n",
       " ILE153 to PHE71,\n",
       " ILE153 to TRP158,\n",
       " ILE153 to VAL126,\n",
       " ILE154 to ILE159,\n",
       " ILE154 to ILE72,\n",
       " ILE154 to LEU75,\n",
       " ILE154 to LYS149,\n",
       " ILE154 to PHE71,\n",
       " ILE154 to SER74,\n",
       " ILE154 to THR123,\n",
       " ILE154 to THR73,\n",
       " ILE154 to TYR70,\n",
       " ILE154 to VAL126,\n",
       " ILE154 to VAL67,\n",
       " ILE159 to LEU115,\n",
       " ILE159 to THR164,\n",
       " ILE169 to LEU115,\n",
       " ILE169 to LEU163,\n",
       " ILE169 to PHE108,\n",
       " ILE169 to PHE193,\n",
       " ILE169 to SER111,\n",
       " ILE169 to THR110,\n",
       " ILE169 to THR164,\n",
       " ILE169 to THR195,\n",
       " ILE169 to TYR174,\n",
       " ILE169 to TYR199,\n",
       " ILE169 to VAL114,\n",
       " ILE182 to PHE194,\n",
       " ILE182 to THR177,\n",
       " ILE201 to ILE294,\n",
       " ILE201 to ILE298,\n",
       " ILE201 to THR195,\n",
       " ILE201 to TYR209,\n",
       " ILE201 to VAL206,\n",
       " ILE201 to VAL297,\n",
       " ILE205 to ILE294,\n",
       " ILE205 to LEU212,\n",
       " ILE205 to PHE290,\n",
       " ILE205 to PRO211,\n",
       " ILE205 to TYR199,\n",
       " ILE205 to VAL210,\n",
       " ILE214 to SER220,\n",
       " ILE214 to TYR209,\n",
       " ILE214 to TYR219,\n",
       " ILE214 to VAL126,\n",
       " ILE214 to VAL129,\n",
       " ILE277 to ILE325,\n",
       " ILE277 to LEU324,\n",
       " ILE277 to LYS270,\n",
       " ILE277 to PHE223,\n",
       " ILE277 to PHE282,\n",
       " ILE277 to TYR219,\n",
       " ILE278 to ILE325,\n",
       " ILE278 to ILE72,\n",
       " ILE278 to LEU124,\n",
       " ILE278 to LEU75,\n",
       " ILE278 to LYS273,\n",
       " ILE278 to MET215,\n",
       " ILE278 to PHE321,\n",
       " ILE278 to PRO323,\n",
       " ILE278 to THR283,\n",
       " ILE278 to TYR219,\n",
       " ILE278 to TYR326,\n",
       " ILE291 to ILE298,\n",
       " ILE291 to ILE303,\n",
       " ILE291 to PHE208,\n",
       " ILE291 to TRP286,\n",
       " ILE291 to TYR209,\n",
       " ILE294 to ILE303,\n",
       " ILE294 to PHE289,\n",
       " ILE294 to SER204,\n",
       " ILE294 to TYR209,\n",
       " ILE303 to LEU311,\n",
       " ILE303 to PRO288,\n",
       " ILE303 to TYR308,\n",
       " ILE303 to VAL292,\n",
       " ILE303 to VAL295,\n",
       " ILE303 to VAL297,\n",
       " ILE309 to ILE314,\n",
       " ILE309 to ILE94,\n",
       " ILE309 to MET40,\n",
       " ILE309 to PHE289,\n",
       " ILE309 to TYR316,\n",
       " ILE314 to LEU284,\n",
       " ILE314 to PHE289,\n",
       " ILE314 to PHE321,\n",
       " ILE314 to PRO288,\n",
       " ILE314 to SER319,\n",
       " ILE314 to THR281,\n",
       " ILE314 to TRP286,\n",
       " ILE325 to LEU275,\n",
       " ILE325 to LYS273,\n",
       " ILE325 to MET279,\n",
       " ILE325 to PHE282,\n",
       " ILE325 to THR274,\n",
       " ILE325 to THR281,\n",
       " ILE334 to LEU339,\n",
       " ILE334 to PHE61,\n",
       " ILE334 to SER329,\n",
       " ILE38 to ILE43,\n",
       " ILE38 to LEU95,\n",
       " ILE38 to VAL33,\n",
       " ILE43 to MET36,\n",
       " ILE43 to PHE49,\n",
       " ILE43 to TRP313,\n",
       " ILE43 to TYR316,\n",
       " ILE43 to VAL317,\n",
       " ILE43 to VAL48,\n",
       " ILE43 to VAL87,\n",
       " ILE47 to LEU42,\n",
       " ILE47 to LEU80,\n",
       " ILE47 to LEU84,\n",
       " ILE47 to MET82,\n",
       " ILE47 to PHE321,\n",
       " ILE47 to PRO323,\n",
       " ILE47 to SER319,\n",
       " ILE47 to TYR316,\n",
       " ILE47 to VAL317,\n",
       " ILE47 to VAL52,\n",
       " ILE47 to VAL81,\n",
       " ILE47 to VAL87,\n",
       " ILE55 to ILE72,\n",
       " ILE55 to LEU339,\n",
       " ILE55 to LEU75,\n",
       " ILE55 to LEU80,\n",
       " ILE55 to LYS60,\n",
       " ILE55 to PHE332,\n",
       " ILE55 to PHE336,\n",
       " ILE55 to SER74,\n",
       " ILE55 to THR73,\n",
       " ILE58 to ILE72,\n",
       " ILE58 to LEU339,\n",
       " ILE58 to LEU53,\n",
       " ILE58 to LEU64,\n",
       " ILE58 to LYS147,\n",
       " ILE58 to PHE332,\n",
       " ILE58 to PHE336,\n",
       " ILE58 to THR66,\n",
       " ILE58 to THR68,\n",
       " ILE58 to THR73,\n",
       " ILE58 to TYR326,\n",
       " ILE58 to TYR70,\n",
       " ILE72 to LEU124,\n",
       " ILE72 to LEU275,\n",
       " ILE72 to PHE332,\n",
       " ILE72 to PRO323,\n",
       " ILE72 to THR123,\n",
       " ILE72 to THR66,\n",
       " ILE72 to TYR326,\n",
       " ILE72 to VAL54,\n",
       " ILE72 to VAL67,\n",
       " ILE94 to LYS305,\n",
       " ILE94 to MET36,\n",
       " ILE94 to MET40,\n",
       " ILE94 to PHE89,\n",
       " ILE94 to TRP313,\n",
       " ILE94 to TRP32,\n",
       " ILE94 to VAL33,\n",
       " LEU115 to LEU163,\n",
       " LEU115 to PHE166,\n",
       " LEU115 to SER120,\n",
       " LEU115 to SER161,\n",
       " LEU115 to SER165,\n",
       " LEU115 to SER207,\n",
       " LEU115 to THR110,\n",
       " LEU115 to THR164,\n",
       " LEU115 to TRP158,\n",
       " LEU115 to VAL157,\n",
       " LEU115 to VAL160,\n",
       " LEU115 to VAL81,\n",
       " LEU124 to LEU275,\n",
       " LEU124 to LEU75,\n",
       " LEU124 to MET215,\n",
       " LEU124 to MET279,\n",
       " LEU124 to PHE282,\n",
       " LEU124 to PHE71,\n",
       " LEU124 to PRO211,\n",
       " LEU124 to THR283,\n",
       " LEU124 to TRP286,\n",
       " LEU124 to TYR219,\n",
       " LEU124 to TYR326,\n",
       " LEU124 to VAL129,\n",
       " LEU144 to LYS149,\n",
       " LEU144 to PHE133,\n",
       " LEU144 to PHE139,\n",
       " LEU144 to PRO138,\n",
       " LEU144 to SER137,\n",
       " LEU144 to THR136,\n",
       " LEU145 to LYS140,\n",
       " LEU145 to PHE133,\n",
       " LEU145 to PHE71,\n",
       " LEU145 to VAL126,\n",
       " LEU145 to VAL129,\n",
       " LEU145 to VAL67,\n",
       " LEU155 to SER74,\n",
       " LEU155 to TYR70,\n",
       " LEU155 to VAL160,\n",
       " LEU163 to PRO168,\n",
       " LEU163 to TRP158,\n",
       " LEU163 to TRP173,\n",
       " LEU163 to TYR199,\n",
       " LEU167 to SER111,\n",
       " LEU167 to SER161,\n",
       " LEU167 to TRP173,\n",
       " LEU167 to TYR199,\n",
       " LEU212 to LEU287,\n",
       " LEU212 to MET279,\n",
       " LEU212 to PHE217,\n",
       " LEU212 to PHE282,\n",
       " LEU212 to SER207,\n",
       " LEU212 to THR283,\n",
       " LEU266 to TYR141,\n",
       " LEU275 to LYS270,\n",
       " LEU275 to PHE223,\n",
       " LEU275 to SER220,\n",
       " LEU275 to THR68,\n",
       " LEU275 to TYR219,\n",
       " LEU275 to TYR326,\n",
       " LEU275 to VAL222,\n",
       " LEU284 to MET279,\n",
       " LEU284 to PHE289,\n",
       " LEU284 to PHE321,\n",
       " LEU287 to LEU311,\n",
       " LEU287 to PHE208,\n",
       " LEU287 to PHE282,\n",
       " LEU287 to TYR209,\n",
       " LEU287 to VAL292,\n",
       " LEU302 to VAL295,\n",
       " LEU302 to VAL307,\n",
       " LEU310 to LYS305,\n",
       " LEU310 to PHE289,\n",
       " LEU310 to TYR316,\n",
       " LEU311 to LYS305,\n",
       " LEU311 to PHE289,\n",
       " LEU311 to PHE290,\n",
       " LEU311 to PRO288,\n",
       " LEU311 to TRP286,\n",
       " LEU311 to TYR316,\n",
       " LEU311 to VAL292,\n",
       " LEU324 to PHE336,\n",
       " LEU324 to VAL54,\n",
       " LEU339 to LEU53,\n",
       " LEU339 to LYS60,\n",
       " LEU339 to PHE332,\n",
       " LEU339 to PHE61,\n",
       " LEU339 to THR56,\n",
       " LEU339 to VAL52,\n",
       " LEU339 to VAL54,\n",
       " LEU340 to LEU53,\n",
       " LEU340 to PHE49,\n",
       " LEU342 to PHE336,\n",
       " LEU45 to MET40,\n",
       " LEU53 to PHE332,\n",
       " LEU53 to PHE336,\n",
       " LEU53 to PRO323,\n",
       " LEU53 to VAL48,\n",
       " LEU64 to PHE332,\n",
       " LEU64 to PHE336,\n",
       " LEU64 to PRO330,\n",
       " LEU64 to SER329,\n",
       " LEU64 to THR56,\n",
       " LEU64 to TYR70,\n",
       " LEU64 to VAL54,\n",
       " LEU75 to LEU80,\n",
       " LEU75 to PHE282,\n",
       " LEU75 to PRO323,\n",
       " LEU75 to SER120,\n",
       " LEU75 to SER319,\n",
       " LEU75 to THR123,\n",
       " LEU75 to TRP158,\n",
       " LEU75 to TYR326,\n",
       " LEU75 to TYR70,\n",
       " LEU80 to PHE49,\n",
       " LEU80 to SER319,\n",
       " LEU80 to VAL48,\n",
       " LEU80 to VAL52,\n",
       " LEU80 to VAL86,\n",
       " LEU84 to PHE89,\n",
       " LEU84 to TRP105,\n",
       " LEU84 to VAL44,\n",
       " LEU84 to VAL48,\n",
       " LEU84 to VAL52,\n",
       " LEU95 to MET36,\n",
       " LEU95 to MET40,\n",
       " LEU95 to SER41,\n",
       " LEU95 to TRP32,\n",
       " LEU95 to VAL33,\n",
       " LEU95 to VAL34,\n",
       " LYS140 to PHE133,\n",
       " LYS147 to THR66,\n",
       " LYS147 to THR68,\n",
       " LYS147 to THR73,\n",
       " LYS147 to TYR70,\n",
       " LYS147 to VAL152,\n",
       " LYS147 to VAL67,\n",
       " LYS149 to PHE71,\n",
       " LYS149 to SER143,\n",
       " LYS227 to VAL222,\n",
       " LYS267 to PRO330,\n",
       " LYS273 to PHE223,\n",
       " LYS273 to TYR219,\n",
       " LYS305 to PHE193,\n",
       " LYS305 to PHE194,\n",
       " LYS97 to THR189,\n",
       " LYS97 to TRP32,\n",
       " LYS97 to VAL33,\n",
       " MET156 to SER161,\n",
       " MET171 to PHE166,\n",
       " MET171 to SER165,\n",
       " MET215 to MET279,\n",
       " MET215 to PHE208,\n",
       " MET215 to PHE282,\n",
       " ...]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_residues"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import custom_tica\n",
    "reload(custom_tica)\n",
    "from custom_tica import *\n",
    "#transform_to_h5(projection_operator_dir, features_dir, projected_features_dir, worker_pool=None, parallel=True, normalizer=compat_verboseload(\"%s/normalizer.h5\" %features_dir))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "tica_object = compat_verboseload(projection_operator_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "with open(projection_operator_dir, 'rb') as f:\n",
    "    tica_object = pickle.load(f, encoding='latin1') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#all_tica = compat_verboseload(projected_features_dir)\n",
    "#all_tica = [x[:,range(1,10)] for x in all_tica]\n",
    "#print(all_tica[0].shape)\n",
    "#print(len(all_tica))\n",
    "#verbosedump(all_tica, projected_features_dir)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#n_components=9"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 737.05920774,  720.14166836,  669.30667022,  579.91041367,\n",
       "        445.26316753,  398.429576  ,  397.17247717,  390.47198856,\n",
       "        372.69426165,  296.83517177])"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tica_object.timescales_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loading \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/phi_psi_chi2_allprot_projected.h5\"...\n",
      "()\n",
      "None\n",
      "Done plotting columns\n"
     ]
    }
   ],
   "source": [
    "plot_columns(tica_dir, projected_features_dir, titles = [\"tIC%d\" %j for j in range(1,11)], tICA = True, scale = 1.0, refcoords_file = None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "feature_indices = []\n",
    "traj_basenames = get_trajectory_files(traj_dir, \".h5\")\n",
    "traj_basenames = [os.path.splitext(os.path.basename(t))[0] for t in traj_basenames]\n",
    "feature_basenames = get_trajectory_files(features_dir, \".dataset\")\n",
    "feature_basenames = [os.path.splitext(os.path.basename(t))[0] for t in feature_basenames]\n",
    "for i, traj_basename in enumerate(traj_basenames):\n",
    "    if traj_basename in feature_basenames:\n",
    "        feature_indices.append(int(i))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "tica_coords = compat_verboseload(projected_features_dir)\n",
    "#pnas_coords = compat_verboseload(user_defined_features_file)\n",
    "#pnas_coords = [pnas_coords[i] for i in feature_indices]\n",
    "tica_names = [\"tIC.%d\" %i for i in range(1,n_components+1)]\n",
    "#pnas_names = sorted(feature_name_residues_dict.keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(139932, 10)\n"
     ]
    }
   ],
   "source": [
    "# prints total number of frames that we have (t=0 - t=n)\n",
    "print(np.concatenate(tica_coords).shape)\n",
    "#print(np.concatenate(pnas_coords).shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from plots import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#plot_data_vs_data(np.concatenate(tica_coords), np.concatenate(pnas_coords), tica_names, pnas_names, analysis_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "subsampled_features_dir = os.path.join(tica_dir, \"subsampled_features\")\n",
    "if not os.path.exists(subsampled_features_dir): os.makedirs(subsampled_features_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "feature_names = compat_verboseload(feature_residues_pkl)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loading /home/enf/md_simulations/b2ar/Salmeterol/featurescontact_6pt6_angstroms-CA-py3-far/reference_receptors/2RH1_prepped-A.dataset\n",
      "loading /home/enf/md_simulations/b2ar/Salmeterol/featurescontact_6pt6_angstroms-CA-py3-far/reference_receptors/3P0G_pymol_prepped-A.dataset\n",
      "loading /home/enf/md_simulations/b2ar/Salmeterol/featurescontact_6pt6_angstroms-CA-py3-far/reference_receptors/ionized.dataset\n",
      "3\n",
      "(1, 2610)\n",
      "(1, 2610)\n",
      "2610\n"
     ]
    }
   ],
   "source": [
    "ref_features = [load_file(f) for f in get_trajectory_files(ref_features_dir, \".dataset\")]\n",
    "print(len(ref_features))\n",
    "print(ref_features[0].shape)\n",
    "print(ref_features[1].shape)\n",
    "print(len(feature_names))\n",
    "ref_features = pd.DataFrame(np.vstack(ref_features), columns=[str(f) for f in feature_names])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ALA119 to ALA78</th>\n",
       "      <th>ALA119 to ASP79</th>\n",
       "      <th>ALA119 to LEU124</th>\n",
       "      <th>ALA119 to LEU75</th>\n",
       "      <th>ALA119 to MET82</th>\n",
       "      <th>ALA119 to SER161</th>\n",
       "      <th>ALA119 to SER207</th>\n",
       "      <th>ALA119 to SER74</th>\n",
       "      <th>ALA119 to TRP158</th>\n",
       "      <th>ALA119 to VAL114</th>\n",
       "      <th>...</th>\n",
       "      <th>VAL292_CA to VAL297_CA</th>\n",
       "      <th>VAL292_CA to VAL307_CA</th>\n",
       "      <th>VAL317_CA to VAL87_CA</th>\n",
       "      <th>VAL34_CA to VAL39_CA</th>\n",
       "      <th>VAL39_CA to VAL44_CA</th>\n",
       "      <th>VAL44_CA to VAL86_CA</th>\n",
       "      <th>VAL44_CA to VAL87_CA</th>\n",
       "      <th>VAL48_CA to VAL81_CA</th>\n",
       "      <th>VAL81_CA to VAL86_CA</th>\n",
       "      <th>VAL81_CA to VAL87_CA</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.461015</td>\n",
       "      <td>0.776682</td>\n",
       "      <td>0.486704</td>\n",
       "      <td>0.526418</td>\n",
       "      <td>0.894512</td>\n",
       "      <td>0.325657</td>\n",
       "      <td>0.624341</td>\n",
       "      <td>0.559677</td>\n",
       "      <td>0.379954</td>\n",
       "      <td>0.494914</td>\n",
       "      <td>...</td>\n",
       "      <td>0.856652</td>\n",
       "      <td>0.969607</td>\n",
       "      <td>0.871400</td>\n",
       "      <td>0.855769</td>\n",
       "      <td>0.883398</td>\n",
       "      <td>0.954805</td>\n",
       "      <td>0.663171</td>\n",
       "      <td>0.835061</td>\n",
       "      <td>0.688208</td>\n",
       "      <td>0.882706</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.376533</td>\n",
       "      <td>0.656476</td>\n",
       "      <td>0.488533</td>\n",
       "      <td>0.495369</td>\n",
       "      <td>0.652861</td>\n",
       "      <td>0.324367</td>\n",
       "      <td>0.638784</td>\n",
       "      <td>0.520898</td>\n",
       "      <td>0.355660</td>\n",
       "      <td>0.475796</td>\n",
       "      <td>...</td>\n",
       "      <td>0.847109</td>\n",
       "      <td>0.907385</td>\n",
       "      <td>0.924539</td>\n",
       "      <td>0.859190</td>\n",
       "      <td>0.883134</td>\n",
       "      <td>0.934655</td>\n",
       "      <td>0.626340</td>\n",
       "      <td>0.816161</td>\n",
       "      <td>0.699942</td>\n",
       "      <td>0.874036</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.394422</td>\n",
       "      <td>0.679888</td>\n",
       "      <td>0.481039</td>\n",
       "      <td>0.518173</td>\n",
       "      <td>0.811594</td>\n",
       "      <td>0.314504</td>\n",
       "      <td>0.644040</td>\n",
       "      <td>0.546482</td>\n",
       "      <td>0.383042</td>\n",
       "      <td>0.479343</td>\n",
       "      <td>...</td>\n",
       "      <td>0.860910</td>\n",
       "      <td>0.847107</td>\n",
       "      <td>0.928936</td>\n",
       "      <td>0.880690</td>\n",
       "      <td>0.872936</td>\n",
       "      <td>0.951248</td>\n",
       "      <td>0.652403</td>\n",
       "      <td>0.829677</td>\n",
       "      <td>0.711434</td>\n",
       "      <td>0.901588</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 2610 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   ALA119 to ALA78  ALA119 to ASP79  ALA119 to LEU124  ALA119 to LEU75  \\\n",
       "0         0.461015         0.776682          0.486704         0.526418   \n",
       "1         0.376533         0.656476          0.488533         0.495369   \n",
       "2         0.394422         0.679888          0.481039         0.518173   \n",
       "\n",
       "   ALA119 to MET82  ALA119 to SER161  ALA119 to SER207  ALA119 to SER74  \\\n",
       "0         0.894512          0.325657          0.624341         0.559677   \n",
       "1         0.652861          0.324367          0.638784         0.520898   \n",
       "2         0.811594          0.314504          0.644040         0.546482   \n",
       "\n",
       "   ALA119 to TRP158  ALA119 to VAL114          ...           \\\n",
       "0          0.379954          0.494914          ...            \n",
       "1          0.355660          0.475796          ...            \n",
       "2          0.383042          0.479343          ...            \n",
       "\n",
       "   VAL292_CA to VAL297_CA  VAL292_CA to VAL307_CA  VAL317_CA to VAL87_CA  \\\n",
       "0                0.856652                0.969607               0.871400   \n",
       "1                0.847109                0.907385               0.924539   \n",
       "2                0.860910                0.847107               0.928936   \n",
       "\n",
       "   VAL34_CA to VAL39_CA  VAL39_CA to VAL44_CA  VAL44_CA to VAL86_CA  \\\n",
       "0              0.855769              0.883398              0.954805   \n",
       "1              0.859190              0.883134              0.934655   \n",
       "2              0.880690              0.872936              0.951248   \n",
       "\n",
       "   VAL44_CA to VAL87_CA  VAL48_CA to VAL81_CA  VAL81_CA to VAL86_CA  \\\n",
       "0              0.663171              0.835061              0.688208   \n",
       "1              0.626340              0.816161              0.699942   \n",
       "2              0.652403              0.829677              0.711434   \n",
       "\n",
       "   VAL81_CA to VAL87_CA  \n",
       "0              0.882706  \n",
       "1              0.874036  \n",
       "2              0.901588  \n",
       "\n",
       "[3 rows x 2610 columns]"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ref_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Interpreting tIC 1\n",
      "feature_importances_df.shape\n",
      "(12, 5)\n",
      "residue_importances_df.shape\n",
      "(18, 2)\n",
      "              feature_name      res_i      res_j  importance  \\\n",
      "10  Gly102_Ca To Thr189_Ca  Gly102_Ca  Thr189_Ca    0.570747   \n",
      "4         Gly102 To Thr189     Gly102     Thr189   -0.396816   \n",
      "1         Asn103 To Cys191     Asn103     Cys191   -0.342891   \n",
      "5   Asn103_Ca To Glu188_Ca  Asn103_Ca  Glu188_Ca    0.338240   \n",
      "11   Lys97_Ca To Thr189_Ca   Lys97_Ca  Thr189_Ca    0.261177   \n",
      "9   Glu188_Ca To Phe104_Ca  Glu188_Ca  Phe104_Ca   -0.242669   \n",
      "3         Glu188 To Thr100     Glu188     Thr100    0.242470   \n",
      "8   Glu187_Ca To Gly102_Ca  Glu187_Ca  Gly102_Ca   -0.161114   \n",
      "0          Ala76 To Cys327      Ala76     Cys327    0.149930   \n",
      "7   Cys190_Ca To Phe101_Ca  Cys190_Ca  Phe101_Ca    0.121625   \n",
      "6   Cys106_Ca To Cys190_Ca  Cys106_Ca  Cys190_Ca   -0.082715   \n",
      "2         Asn103 To Glu188     Asn103     Glu188    0.061434   \n",
      "\n",
      "                   feature  \n",
      "10  GLY102_CA to THR189_CA  \n",
      "4         GLY102 to THR189  \n",
      "1         ASN103 to CYS191  \n",
      "5   ASN103_CA to GLU188_CA  \n",
      "11   LYS97_CA to THR189_CA  \n",
      "9   GLU188_CA to PHE104_CA  \n",
      "3         GLU188 to THR100  \n",
      "8   GLU187_CA to GLY102_CA  \n",
      "0          ALA76 to CYS327  \n",
      "7   CYS190_CA to PHE101_CA  \n",
      "6   CYS106_CA to CYS190_CA  \n",
      "2         ASN103 to GLU188  \n",
      "Using dark_background\n",
      "             residue  importance\n",
      "Thr189_Ca  Thr189_Ca    0.555269\n",
      "Gly102_Ca  Gly102_Ca    0.534154\n",
      "Gly102        Gly102   -0.396816\n",
      "Thr189        Thr189   -0.396816\n",
      "Cys191        Cys191   -0.342891\n",
      "Asn103_Ca  Asn103_Ca    0.338240\n",
      "Glu188_Ca  Glu188_Ca    0.309194\n",
      "Lys97_Ca    Lys97_Ca    0.261177\n",
      "Phe104_Ca  Phe104_Ca   -0.242669\n",
      "Thr100        Thr100    0.242470\n",
      "Glu188        Glu188    0.233418\n",
      "Glu187_Ca  Glu187_Ca   -0.161114\n",
      "Cys327        Cys327    0.149930\n",
      "Ala76          Ala76    0.149930\n",
      "Phe101_Ca  Phe101_Ca    0.121625\n",
      "Cys190_Ca  Cys190_Ca    0.111408\n",
      "Cys106_Ca  Cys106_Ca   -0.082715\n",
      "Asn103        Asn103    0.041218\n",
      "Using dark_background\n",
      "Interpreting tIC 2\n",
      "feature_importances_df.shape\n",
      "(13, 5)\n",
      "residue_importances_df.shape\n",
      "(21, 2)\n",
      "              feature_name      res_i      res_j  importance  \\\n",
      "5          Asp79 To Gly320      Asp79     Gly320   -0.317969   \n",
      "10   Leu75_Ca To Pro323_Ca   Leu75_Ca  Pro323_Ca   -0.227773   \n",
      "1          Ala46 To Pro323      Ala46     Pro323    0.211151   \n",
      "7    Cys327_Ca To Ile72_Ca  Cys327_Ca   Ile72_Ca    0.184326   \n",
      "9    Ile72_Ca To Pro323_Ca   Ile72_Ca  Pro323_Ca   -0.153175   \n",
      "4          Asn51 To Gly320      Asn51     Gly320   -0.138799   \n",
      "6          Gly50 To Ser319      Gly50     Ser319    0.135587   \n",
      "8   Cys327_Ca To Thr274_Ca  Cys327_Ca  Thr274_Ca   -0.134291   \n",
      "11   Leu75_Ca To Ser319_Ca   Leu75_Ca  Ser319_Ca    0.133830   \n",
      "3         Asn322 To Ser120     Asn322     Ser120    0.109498   \n",
      "2         Asn103 To Phe108     Asn103     Phe108    0.105721   \n",
      "0         Ala134 To Lys267     Ala134     Lys267    0.071751   \n",
      "12   Thr189_Ca To Trp99_Ca  Thr189_Ca   Trp99_Ca    0.063104   \n",
      "\n",
      "                   feature  \n",
      "5          ASP79 to GLY320  \n",
      "10   LEU75_CA to PRO323_CA  \n",
      "1          ALA46 to PRO323  \n",
      "7    CYS327_CA to ILE72_CA  \n",
      "9    ILE72_CA to PRO323_CA  \n",
      "4          ASN51 to GLY320  \n",
      "6          GLY50 to SER319  \n",
      "8   CYS327_CA to THR274_CA  \n",
      "11   LEU75_CA to SER319_CA  \n",
      "3         ASN322 to SER120  \n",
      "2         ASN103 to PHE108  \n",
      "0         ALA134 to LYS267  \n",
      "12   THR189_CA to TRP99_CA  \n",
      "Using dark_background\n",
      "             residue  importance\n",
      "Asp79          Asp79   -0.317969\n",
      "Ala46          Ala46    0.211151\n",
      "Pro323        Pro323    0.211151\n",
      "Cys327_Ca  Cys327_Ca    0.168395\n",
      "Ile72_Ca    Ile72_Ca    0.167451\n",
      "Pro323_Ca  Pro323_Ca   -0.156905\n",
      "Gly320        Gly320   -0.147757\n",
      "Asn51          Asn51   -0.138799\n",
      "Ser319        Ser319    0.135587\n",
      "Gly50          Gly50    0.135587\n",
      "Thr274_Ca  Thr274_Ca   -0.134291\n",
      "Ser319_Ca  Ser319_Ca    0.133830\n",
      "Leu75_Ca    Leu75_Ca    0.115749\n",
      "Ser120        Ser120    0.109498\n",
      "Asn322        Asn322    0.109498\n",
      "Phe108        Phe108    0.105721\n",
      "Asn103        Asn103    0.105721\n",
      "Lys267        Lys267    0.071751\n",
      "Ala134        Ala134    0.071751\n",
      "Thr189_Ca  Thr189_Ca    0.063104\n",
      "Trp99_Ca    Trp99_Ca    0.063104\n",
      "Using dark_background\n",
      "Interpreting tIC 3\n",
      "feature_importances_df.shape\n",
      "(15, 5)\n",
      "residue_importances_df.shape\n",
      "(29, 2)\n",
      "              feature_name      res_i      res_j  importance  \\\n",
      "4         Asp130 To Tyr141     Asp130     Tyr141    0.366446   \n",
      "3         Asn103 To Phe108     Asn103     Phe108    0.180801   \n",
      "2         Arg328 To Lys273     Arg328     Lys273    0.169284   \n",
      "11  Ile278_Ca To Phe321_Ca  Ile278_Ca  Phe321_Ca   -0.165104   \n",
      "7          Ser74 To Trp158      Ser74     Trp158    0.156094   \n",
      "9   Asn103_Ca To Cys190_Ca  Asn103_Ca  Cys190_Ca    0.147722   \n",
      "12  Ser120_Ca To Ser319_Ca  Ser120_Ca  Ser319_Ca   -0.144887   \n",
      "13   Thr189_Ca To Trp99_Ca  Thr189_Ca   Trp99_Ca    0.141763   \n",
      "14   Trp313_Ca To Val39_Ca  Trp313_Ca   Val39_Ca   -0.128684   \n",
      "0         Ala134 To Lys267     Ala134     Lys267    0.115878   \n",
      "8          Tyr70 To Val152      Tyr70     Val152   -0.114278   \n",
      "5          Gly320 To Ile43     Gly320      Ile43   -0.106877   \n",
      "1          Arg131 To Thr68     Arg131      Thr68    0.101924   \n",
      "6          Ser74 To Thr123      Ser74     Thr123    0.086680   \n",
      "10  Asp130_Ca To Ser143_Ca  Asp130_Ca  Ser143_Ca   -0.067476   \n",
      "\n",
      "                   feature  \n",
      "4         ASP130 to TYR141  \n",
      "3         ASN103 to PHE108  \n",
      "2         ARG328 to LYS273  \n",
      "11  ILE278_CA to PHE321_CA  \n",
      "7          SER74 to TRP158  \n",
      "9   ASN103_CA to CYS190_CA  \n",
      "12  SER120_CA to SER319_CA  \n",
      "13   THR189_CA to TRP99_CA  \n",
      "14   TRP313_CA to VAL39_CA  \n",
      "0         ALA134 to LYS267  \n",
      "8          TYR70 to VAL152  \n",
      "5          GLY320 to ILE43  \n",
      "1          ARG131 to THR68  \n",
      "6          SER74 to THR123  \n",
      "10  ASP130_CA to SER143_CA  \n",
      "Using dark_background\n",
      "             residue  importance\n",
      "Asp130        Asp130    0.366446\n",
      "Tyr141        Tyr141    0.366446\n",
      "Asn103        Asn103    0.180801\n",
      "Phe108        Phe108    0.180801\n",
      "Arg328        Arg328    0.169284\n",
      "Lys273        Lys273    0.169284\n",
      "Phe321_Ca  Phe321_Ca   -0.165104\n",
      "Ile278_Ca  Ile278_Ca   -0.165104\n",
      "Trp158        Trp158    0.156094\n",
      "Ser74          Ser74    0.152623\n",
      "Cys190_Ca  Cys190_Ca    0.147722\n",
      "Asn103_Ca  Asn103_Ca    0.147722\n",
      "Ser319_Ca  Ser319_Ca   -0.144887\n",
      "Ser120_Ca  Ser120_Ca   -0.144887\n",
      "Trp99_Ca    Trp99_Ca    0.141763\n",
      "Thr189_Ca  Thr189_Ca    0.141763\n",
      "Val39_Ca    Val39_Ca   -0.128684\n",
      "Trp313_Ca  Trp313_Ca   -0.128684\n",
      "Lys267        Lys267    0.115878\n",
      "Ala134        Ala134    0.115878\n",
      "Tyr70          Tyr70   -0.114278\n",
      "Val152        Val152   -0.114278\n",
      "Ile43          Ile43   -0.106877\n",
      "Gly320        Gly320   -0.106877\n",
      "Thr68          Thr68    0.101924\n",
      "Arg131        Arg131    0.101924\n",
      "Thr123        Thr123    0.086680\n",
      "Asp130_Ca  Asp130_Ca   -0.067476\n",
      "Ser143_Ca  Ser143_Ca   -0.067476\n",
      "Using dark_background\n",
      "Interpreting tIC 4\n",
      "feature_importances_df.shape\n",
      "(15, 5)\n",
      "residue_importances_df.shape\n",
      "(25, 2)\n",
      "              feature_name      res_i      res_j  importance  \\\n",
      "2         Asn103 To Phe108     Asn103     Phe108    0.555565   \n",
      "10  Asp130_Ca To Ser143_Ca  Asp130_Ca  Ser143_Ca    0.457215   \n",
      "4         Thr146 To Tyr141     Thr146     Tyr141   -0.294744   \n",
      "11  Asp130_Ca To Tyr141_Ca  Asp130_Ca  Tyr141_Ca   -0.279399   \n",
      "8   Asn103_Ca To Phe108_Ca  Asn103_Ca  Phe108_Ca   -0.243000   \n",
      "1         Asn103 To Cys191     Asn103     Cys191    0.205201   \n",
      "13  Leu144_Ca To Phe133_Ca  Leu144_Ca  Phe133_Ca   -0.202779   \n",
      "6   Ala134_Ca To Lys140_Ca  Ala134_Ca  Lys140_Ca    0.186717   \n",
      "14  Leu145_Ca To Lys140_Ca  Leu145_Ca  Lys140_Ca    0.184228   \n",
      "3         His269 To Tyr141     His269     Tyr141    0.147901   \n",
      "7   Arg131_Ca To Gln142_Ca  Arg131_Ca  Gln142_Ca   -0.146921   \n",
      "12   Cys190_Ca To His93_Ca  Cys190_Ca   His93_Ca    0.135923   \n",
      "9   Asn318_Ca To Ile121_Ca  Asn318_Ca  Ile121_Ca   -0.111965   \n",
      "5          Thr68 To Tyr141      Thr68     Tyr141   -0.089018   \n",
      "0         Ala181 To Ala186     Ala181     Ala186   -0.067644   \n",
      "\n",
      "                   feature  \n",
      "2         ASN103 to PHE108  \n",
      "10  ASP130_CA to SER143_CA  \n",
      "4         THR146 to TYR141  \n",
      "11  ASP130_CA to TYR141_CA  \n",
      "8   ASN103_CA to PHE108_CA  \n",
      "1         ASN103 to CYS191  \n",
      "13  LEU144_CA to PHE133_CA  \n",
      "6   ALA134_CA to LYS140_CA  \n",
      "14  LEU145_CA to LYS140_CA  \n",
      "3         HIS269 to TYR141  \n",
      "7   ARG131_CA to GLN142_CA  \n",
      "12   CYS190_CA to HIS93_CA  \n",
      "9   ASN318_CA to ILE121_CA  \n",
      "5          THR68 to TYR141  \n",
      "0         ALA181 to ALA186  \n",
      "Using dark_background\n",
      "             residue  importance\n",
      "Phe108        Phe108    0.555565\n",
      "Asn103        Asn103    0.538047\n",
      "Ser143_Ca  Ser143_Ca    0.457215\n",
      "Asp130_Ca  Asp130_Ca    0.420384\n",
      "Thr146        Thr146   -0.294744\n",
      "Tyr141_Ca  Tyr141_Ca   -0.279399\n",
      "Phe108_Ca  Phe108_Ca   -0.243000\n",
      "Asn103_Ca  Asn103_Ca   -0.243000\n",
      "Cys191        Cys191    0.205201\n",
      "Leu144_Ca  Leu144_Ca   -0.202779\n",
      "Phe133_Ca  Phe133_Ca   -0.202779\n",
      "Ala134_Ca  Ala134_Ca    0.186717\n",
      "Lys140_Ca  Lys140_Ca    0.186593\n",
      "Leu145_Ca  Leu145_Ca    0.184228\n",
      "His269        His269    0.147901\n",
      "Gln142_Ca  Gln142_Ca   -0.146921\n",
      "Arg131_Ca  Arg131_Ca   -0.146921\n",
      "Cys190_Ca  Cys190_Ca    0.135923\n",
      "His93_Ca    His93_Ca    0.135923\n",
      "Tyr141        Tyr141    0.124209\n",
      "Asn318_Ca  Asn318_Ca   -0.111965\n",
      "Ile121_Ca  Ile121_Ca   -0.111965\n",
      "Thr68          Thr68   -0.089018\n",
      "Ala186        Ala186   -0.067644\n",
      "Ala181        Ala181   -0.067644\n",
      "Using dark_background\n",
      "Interpreting tIC 5\n",
      "feature_importances_df.shape\n",
      "(26, 5)\n",
      "residue_importances_df.shape\n",
      "(35, 2)\n",
      "              feature_name      res_i      res_j  importance  \\\n",
      "7   Ala134_Ca To Leu144_Ca  Ala134_Ca  Leu144_Ca   -0.685624   \n",
      "17  Leu144_Ca To Thr136_Ca  Leu144_Ca  Thr136_Ca    0.611477   \n",
      "13  Ile135_Ca To Tyr141_Ca  Ile135_Ca  Tyr141_Ca    0.599418   \n",
      "9   Arg131_Ca To Leu145_Ca  Arg131_Ca  Leu145_Ca    0.332531   \n",
      "24  Thr136_Ca To Tyr141_Ca  Thr136_Ca  Tyr141_Ca   -0.331778   \n",
      "12  Ile135_Ca To Lys140_Ca  Ile135_Ca  Lys140_Ca   -0.331752   \n",
      "3         Leu145 To Lys140     Leu145     Lys140   -0.312285   \n",
      "16  Leu144_Ca To Ser137_Ca  Leu144_Ca  Ser137_Ca   -0.308276   \n",
      "15  Leu144_Ca To Phe139_Ca  Leu144_Ca  Phe139_Ca    0.294746   \n",
      "8   Ala134_Ca To Lys140_Ca  Ala134_Ca  Lys140_Ca    0.268427   \n",
      "6         Thr146 To Tyr141     Thr146     Tyr141   -0.262145   \n",
      "23   Phe71_Ca To Tyr141_Ca   Phe71_Ca  Tyr141_Ca    0.241289   \n",
      "14    Ile94_Ca To Trp32_Ca   Ile94_Ca   Trp32_Ca    0.225031   \n",
      "0         Ala134 To Tyr141     Ala134     Tyr141    0.221464   \n",
      "19  Lys140_Ca To Phe133_Ca  Lys140_Ca  Phe133_Ca   -0.214982   \n",
      "18    Leu95_Ca To Trp32_Ca   Leu95_Ca   Trp32_Ca   -0.208498   \n",
      "5         Thr136 To Tyr141     Thr136     Tyr141   -0.185824   \n",
      "22   Phe71_Ca To Ser143_Ca   Phe71_Ca  Ser143_Ca   -0.169888   \n",
      "25   Trp313_Ca To Val39_Ca  Trp313_Ca   Val39_Ca    0.166559   \n",
      "2         Leu144 To Thr136     Leu144     Thr136    0.142547   \n",
      "1          Asp79 To Ser120      Asp79     Ser120    0.123508   \n",
      "10  Arg151_Ca To Leu145_Ca  Arg151_Ca  Leu145_Ca    0.106049   \n",
      "11  Asn322_Ca To Phe282_Ca  Asn322_Ca  Phe282_Ca    0.094351   \n",
      "20  Phe289_Ca To Trp313_Ca  Phe289_Ca  Trp313_Ca    0.093920   \n",
      "4         Phe290 To Tyr209     Phe290     Tyr209    0.078538   \n",
      "21  Phe290_Ca To Ser203_Ca  Phe290_Ca  Ser203_Ca    0.064452   \n",
      "\n",
      "                   feature  \n",
      "7   ALA134_CA to LEU144_CA  \n",
      "17  LEU144_CA to THR136_CA  \n",
      "13  ILE135_CA to TYR141_CA  \n",
      "9   ARG131_CA to LEU145_CA  \n",
      "24  THR136_CA to TYR141_CA  \n",
      "12  ILE135_CA to LYS140_CA  \n",
      "3         LEU145 to LYS140  \n",
      "16  LEU144_CA to SER137_CA  \n",
      "15  LEU144_CA to PHE139_CA  \n",
      "8   ALA134_CA to LYS140_CA  \n",
      "6         THR146 to TYR141  \n",
      "23   PHE71_CA to TYR141_CA  \n",
      "14    ILE94_CA to TRP32_CA  \n",
      "0         ALA134 to TYR141  \n",
      "19  LYS140_CA to PHE133_CA  \n",
      "18    LEU95_CA to TRP32_CA  \n",
      "5         THR136 to TYR141  \n",
      "22   PHE71_CA to SER143_CA  \n",
      "25   TRP313_CA to VAL39_CA  \n",
      "2         LEU144 to THR136  \n",
      "1          ASP79 to SER120  \n",
      "10  ARG151_CA to LEU145_CA  \n",
      "11  ASN322_CA to PHE282_CA  \n",
      "20  PHE289_CA to TRP313_CA  \n",
      "4         PHE290 to TYR209  \n",
      "21  PHE290_CA to SER203_CA  \n",
      "Using dark_background\n",
      "             residue  importance\n",
      "Thr136_Ca  Thr136_Ca    0.564315\n",
      "Leu144_Ca  Leu144_Ca    0.563968\n",
      "Tyr141_Ca  Tyr141_Ca    0.563605\n",
      "Ile135_Ca  Ile135_Ca    0.552860\n",
      "Arg131_Ca  Arg131_Ca    0.332531\n",
      "Leu145_Ca  Leu145_Ca    0.321207\n",
      "Leu145        Leu145   -0.312285\n",
      "Lys140        Lys140   -0.312285\n",
      "Ser137_Ca  Ser137_Ca   -0.308276\n",
      "Phe139_Ca  Phe139_Ca    0.294746\n",
      "Thr146        Thr146   -0.262145\n",
      "Ile94_Ca    Ile94_Ca    0.225031\n",
      "Ala134        Ala134    0.221464\n",
      "Phe71_Ca    Phe71_Ca    0.220730\n",
      "Ala134_Ca  Ala134_Ca    0.220724\n",
      "Lys140_Ca  Lys140_Ca    0.220086\n",
      "Phe133_Ca  Phe133_Ca   -0.214982\n",
      "Leu95_Ca    Leu95_Ca   -0.208498\n",
      "Trp32_Ca    Trp32_Ca    0.203354\n",
      "Tyr141        Tyr141    0.180735\n",
      "Ser143_Ca  Ser143_Ca   -0.169888\n",
      "Val39_Ca    Val39_Ca    0.166559\n",
      "Trp313_Ca  Trp313_Ca    0.162927\n",
      "Leu144        Leu144    0.142547\n",
      "Thr136        Thr136    0.126128\n",
      "Ser120        Ser120    0.123508\n",
      "Asp79          Asp79    0.123508\n",
      "Arg151_Ca  Arg151_Ca    0.106049\n",
      "Phe282_Ca  Phe282_Ca    0.094351\n",
      "Asn322_Ca  Asn322_Ca    0.094351\n",
      "Phe289_Ca  Phe289_Ca    0.093920\n",
      "Tyr209        Tyr209    0.078538\n",
      "Phe290        Phe290    0.078538\n",
      "Phe290_Ca  Phe290_Ca    0.064452\n",
      "Ser203_Ca  Ser203_Ca    0.064452\n",
      "Using dark_background\n",
      "Interpreting tIC 6\n",
      "feature_importances_df.shape\n",
      "(23, 5)\n",
      "residue_importances_df.shape\n",
      "(40, 2)\n",
      "              feature_name      res_i      res_j  importance  \\\n",
      "10   Ala46_Ca To Phe321_Ca   Ala46_Ca  Phe321_Ca    0.376847   \n",
      "0          Ala46 To Phe321      Ala46     Phe321   -0.280282   \n",
      "15  Leu124_Ca To Thr283_Ca  Leu124_Ca  Thr283_Ca    0.243199   \n",
      "7          Pro88 To Trp313      Pro88     Trp313    0.237821   \n",
      "9   Ala226_Ca To Leu275_Ca  Ala226_Ca  Leu275_Ca   -0.232482   \n",
      "4          Gly320 To Ile43     Gly320      Ile43   -0.192746   \n",
      "5         Phe208 To Trp286     Phe208     Trp286    0.164220   \n",
      "18  Leu275_Ca To Ser220_Ca  Leu275_Ca  Ser220_Ca    0.161288   \n",
      "16  Leu124_Ca To Trp286_Ca  Leu124_Ca  Trp286_Ca   -0.156367   \n",
      "2          Asp79 To Trp158      Asp79     Trp158    0.147138   \n",
      "19  Leu311_Ca To Phe290_Ca  Leu311_Ca  Phe290_Ca    0.144484   \n",
      "1         Asp130 To Tyr141     Asp130     Tyr141   -0.125580   \n",
      "8   Ala226_Ca To Glu268_Ca  Ala226_Ca  Glu268_Ca    0.122661   \n",
      "6         Phe289 To Tyr308     Phe289     Tyr308    0.118590   \n",
      "20   Leu324_Ca To Val54_Ca  Leu324_Ca   Val54_Ca    0.115376   \n",
      "11  Asp130_Ca To Leu144_Ca  Asp130_Ca  Leu144_Ca   -0.115342   \n",
      "3         Gly276 To Phe223     Gly276     Phe223    0.114900   \n",
      "12  Cys285_Ca To Leu311_Ca  Cys285_Ca  Leu311_Ca   -0.107696   \n",
      "17  Leu144_Ca To Ser137_Ca  Leu144_Ca  Ser137_Ca    0.107367   \n",
      "14  Ile169_Ca To Thr195_Ca  Ile169_Ca  Thr195_Ca    0.105961   \n",
      "22  Thr164_Ca To Val114_Ca  Thr164_Ca  Val114_Ca   -0.090767   \n",
      "21  Phe336_Ca To Pro323_Ca  Phe336_Ca  Pro323_Ca    0.086195   \n",
      "13  Glu225_Ca To Ser137_Ca  Glu225_Ca  Ser137_Ca    0.084914   \n",
      "\n",
      "                   feature  \n",
      "10   ALA46_CA to PHE321_CA  \n",
      "0          ALA46 to PHE321  \n",
      "15  LEU124_CA to THR283_CA  \n",
      "7          PRO88 to TRP313  \n",
      "9   ALA226_CA to LEU275_CA  \n",
      "4          GLY320 to ILE43  \n",
      "5         PHE208 to TRP286  \n",
      "18  LEU275_CA to SER220_CA  \n",
      "16  LEU124_CA to TRP286_CA  \n",
      "2          ASP79 to TRP158  \n",
      "19  LEU311_CA to PHE290_CA  \n",
      "1         ASP130 to TYR141  \n",
      "8   ALA226_CA to GLU268_CA  \n",
      "6         PHE289 to TYR308  \n",
      "20   LEU324_CA to VAL54_CA  \n",
      "11  ASP130_CA to LEU144_CA  \n",
      "3         GLY276 to PHE223  \n",
      "12  CYS285_CA to LEU311_CA  \n",
      "17  LEU144_CA to SER137_CA  \n",
      "14  ILE169_CA to THR195_CA  \n",
      "22  THR164_CA to VAL114_CA  \n",
      "21  PHE336_CA to PRO323_CA  \n",
      "13  GLU225_CA to SER137_CA  \n",
      "Using dark_background\n",
      "             residue  importance\n",
      "Ala46_Ca    Ala46_Ca    0.376847\n",
      "Phe321_Ca  Phe321_Ca    0.376847\n",
      "Phe321        Phe321   -0.280282\n",
      "Ala46          Ala46   -0.280282\n",
      "Thr283_Ca  Thr283_Ca    0.243199\n",
      "Pro88          Pro88    0.237821\n",
      "Trp313        Trp313    0.237821\n",
      "Leu124_Ca  Leu124_Ca    0.223221\n",
      "Ile43          Ile43   -0.192746\n",
      "Gly320        Gly320   -0.192746\n",
      "Trp286        Trp286    0.164220\n",
      "Phe208        Phe208    0.164220\n",
      "Ser220_Ca  Ser220_Ca    0.161288\n",
      "Trp286_Ca  Trp286_Ca   -0.156367\n",
      "Trp158        Trp158    0.147138\n",
      "Asp79          Asp79    0.147138\n",
      "Phe290_Ca  Phe290_Ca    0.144484\n",
      "Leu275_Ca  Leu275_Ca    0.141599\n",
      "Leu311_Ca  Leu311_Ca    0.131875\n",
      "Tyr141        Tyr141   -0.125580\n",
      "Asp130        Asp130   -0.125580\n",
      "Glu268_Ca  Glu268_Ca    0.122661\n",
      "Phe289        Phe289    0.118590\n",
      "Tyr308        Tyr308    0.118590\n",
      "Val54_Ca    Val54_Ca    0.115376\n",
      "Leu324_Ca  Leu324_Ca    0.115376\n",
      "Asp130_Ca  Asp130_Ca   -0.115342\n",
      "Phe223        Phe223    0.114900\n",
      "Gly276        Gly276    0.114900\n",
      "Cys285_Ca  Cys285_Ca   -0.107696\n",
      "Ser137_Ca  Ser137_Ca    0.106244\n",
      "Ile169_Ca  Ile169_Ca    0.105961\n",
      "Thr195_Ca  Thr195_Ca    0.105961\n",
      "Ala226_Ca  Ala226_Ca    0.104904\n",
      "Leu144_Ca  Leu144_Ca    0.096232\n",
      "Thr164_Ca  Thr164_Ca   -0.090767\n",
      "Val114_Ca  Val114_Ca   -0.090767\n",
      "Phe336_Ca  Phe336_Ca    0.086195\n",
      "Pro323_Ca  Pro323_Ca    0.086195\n",
      "Glu225_Ca  Glu225_Ca    0.084914\n",
      "Using dark_background\n",
      "Interpreting tIC 7\n",
      "feature_importances_df.shape\n",
      "(29, 5)\n",
      "residue_importances_df.shape\n",
      "(45, 2)\n",
      "              feature_name      res_i      res_j  importance  \\\n",
      "3         Asn103 To Phe108     Asn103     Phe108    0.499593   \n",
      "20  Ile135_Ca To Tyr141_Ca  Ile135_Ca  Tyr141_Ca   -0.375563   \n",
      "9   Ala134_Ca To Lys140_Ca  Ala134_Ca  Lys140_Ca   -0.337197   \n",
      "27   Trp313_Ca To Val39_Ca  Trp313_Ca   Val39_Ca    0.319244   \n",
      "25  Phe133_Ca To Tyr141_Ca  Phe133_Ca  Tyr141_Ca    0.318390   \n",
      "21    Ile94_Ca To Trp32_Ca   Ile94_Ca   Trp32_Ca    0.288901   \n",
      "23    Leu95_Ca To Trp32_Ca   Leu95_Ca   Trp32_Ca   -0.264199   \n",
      "8   Ala134_Ca To Leu144_Ca  Ala134_Ca  Leu144_Ca    0.257043   \n",
      "13  Asn103_Ca To Phe108_Ca  Asn103_Ca  Phe108_Ca   -0.238197   \n",
      "19  Ile135_Ca To Lys140_Ca  Ile135_Ca  Lys140_Ca    0.235580   \n",
      "7         Thr146 To Tyr141     Thr146     Tyr141    0.229767   \n",
      "12  Asn103_Ca To Glu188_Ca  Asn103_Ca  Glu188_Ca    0.213906   \n",
      "2         Asn103 To Cys191     Asn103     Cys191    0.208673   \n",
      "22  Leu145_Ca To Phe133_Ca  Leu145_Ca  Phe133_Ca   -0.180848   \n",
      "6          Pro88 To Trp313      Pro88     Trp313    0.172828   \n",
      "17  Glu188_Ca To Phe104_Ca  Glu188_Ca  Phe104_Ca   -0.161219   \n",
      "28   Tyr326_Ca To Val54_Ca  Tyr326_Ca   Val54_Ca   -0.153350   \n",
      "0         Ala134 To Lys267     Ala134     Lys267   -0.152081   \n",
      "26   Thr68_Ca To Tyr141_Ca   Thr68_Ca  Tyr141_Ca    0.150498   \n",
      "14  Asn322_Ca To Ile278_Ca  Asn322_Ca  Ile278_Ca    0.121508   \n",
      "24  Phe133_Ca To Pro138_Ca  Phe133_Ca  Pro138_Ca   -0.118520   \n",
      "1          Ala134 To Thr68     Ala134      Thr68   -0.113052   \n",
      "18   Gly50_Ca To Pro323_Ca   Gly50_Ca  Pro323_Ca   -0.112940   \n",
      "16  Asp130_Ca To Ser143_Ca  Asp130_Ca  Ser143_Ca   -0.110031   \n",
      "10    Ala91_Ca To Met36_Ca   Ala91_Ca   Met36_Ca   -0.106881   \n",
      "11  Arg151_Ca To Leu145_Ca  Arg151_Ca  Leu145_Ca   -0.088949   \n",
      "15   Asn51_Ca To Phe321_Ca   Asn51_Ca  Phe321_Ca    0.080665   \n",
      "5         Phe101 To Thr189     Phe101     Thr189   -0.079146   \n",
      "4           Ile38 To Val33      Ile38      Val33   -0.075788   \n",
      "\n",
      "                   feature  \n",
      "3         ASN103 to PHE108  \n",
      "20  ILE135_CA to TYR141_CA  \n",
      "9   ALA134_CA to LYS140_CA  \n",
      "27   TRP313_CA to VAL39_CA  \n",
      "25  PHE133_CA to TYR141_CA  \n",
      "21    ILE94_CA to TRP32_CA  \n",
      "23    LEU95_CA to TRP32_CA  \n",
      "8   ALA134_CA to LEU144_CA  \n",
      "13  ASN103_CA to PHE108_CA  \n",
      "19  ILE135_CA to LYS140_CA  \n",
      "7         THR146 to TYR141  \n",
      "12  ASN103_CA to GLU188_CA  \n",
      "2         ASN103 to CYS191  \n",
      "22  LEU145_CA to PHE133_CA  \n",
      "6          PRO88 to TRP313  \n",
      "17  GLU188_CA to PHE104_CA  \n",
      "28   TYR326_CA to VAL54_CA  \n",
      "0         ALA134 to LYS267  \n",
      "26   THR68_CA to TYR141_CA  \n",
      "14  ASN322_CA to ILE278_CA  \n",
      "24  PHE133_CA to PRO138_CA  \n",
      "1          ALA134 to THR68  \n",
      "18   GLY50_CA to PRO323_CA  \n",
      "16  ASP130_CA to SER143_CA  \n",
      "10    ALA91_CA to MET36_CA  \n",
      "11  ARG151_CA to LEU145_CA  \n",
      "15   ASN51_CA to PHE321_CA  \n",
      "5         PHE101 to THR189  \n",
      "4           ILE38 to VAL33  \n",
      "Using dark_background\n",
      "             residue  importance\n",
      "Phe108        Phe108    0.499593\n",
      "Asn103        Asn103    0.485047\n",
      "Val39_Ca    Val39_Ca    0.319244\n",
      "Trp313_Ca  Trp313_Ca    0.319244\n",
      "Tyr141_Ca  Tyr141_Ca    0.301600\n",
      "Ile94_Ca    Ile94_Ca    0.288901\n",
      "Phe133_Ca  Phe133_Ca    0.274699\n",
      "Leu95_Ca    Leu95_Ca   -0.264199\n",
      "Trp32_Ca    Trp32_Ca    0.261246\n",
      "Leu144_Ca  Leu144_Ca    0.257043\n",
      "Phe108_Ca  Phe108_Ca   -0.238197\n",
      "Thr146        Thr146    0.229767\n",
      "Tyr141        Tyr141    0.229767\n",
      "Ala134_Ca  Ala134_Ca    0.227331\n",
      "Cys191        Cys191    0.208673\n",
      "Lys140_Ca  Lys140_Ca    0.206941\n",
      "Ile135_Ca  Ile135_Ca    0.205023\n",
      "Glu188_Ca  Glu188_Ca    0.195150\n",
      "Asn103_Ca  Asn103_Ca    0.191301\n",
      "Trp313        Trp313    0.172828\n",
      "Pro88          Pro88    0.172828\n",
      "Phe104_Ca  Phe104_Ca   -0.161219\n",
      "Val54_Ca    Val54_Ca   -0.153350\n",
      "Tyr326_Ca  Tyr326_Ca   -0.153350\n",
      "Lys267        Lys267   -0.152081\n",
      "Thr68_Ca    Thr68_Ca    0.150498\n",
      "Ile278_Ca  Ile278_Ca    0.121508\n",
      "Asn322_Ca  Asn322_Ca    0.121508\n",
      "Pro138_Ca  Pro138_Ca   -0.118520\n",
      "Ala134        Ala134   -0.115004\n",
      "Thr68          Thr68   -0.113052\n",
      "Pro323_Ca  Pro323_Ca   -0.112940\n",
      "Gly50_Ca    Gly50_Ca   -0.112940\n",
      "Asp130_Ca  Asp130_Ca   -0.110031\n",
      "Ser143_Ca  Ser143_Ca   -0.110031\n",
      "Ala91_Ca    Ala91_Ca   -0.106881\n",
      "Met36_Ca    Met36_Ca   -0.106881\n",
      "Leu145_Ca  Leu145_Ca   -0.093544\n",
      "Arg151_Ca  Arg151_Ca   -0.088949\n",
      "Asn51_Ca    Asn51_Ca    0.080665\n",
      "Phe321_Ca  Phe321_Ca    0.080665\n",
      "Thr189        Thr189   -0.079146\n",
      "Phe101        Phe101   -0.079146\n",
      "Val33          Val33   -0.075788\n",
      "Ile38          Ile38   -0.075788\n",
      "Using dark_background\n",
      "Interpreting tIC 8\n",
      "feature_importances_df.shape\n",
      "(10, 5)\n",
      "residue_importances_df.shape\n",
      "(15, 2)\n",
      "             feature_name      res_i      res_j  importance  \\\n",
      "2         Asp79 To Ser120      Asp79     Ser120    0.408458   \n",
      "5         Ile47 To Ser319      Ile47     Ser319    0.332933   \n",
      "7   Asn51_Ca To Gly320_Ca   Asn51_Ca  Gly320_Ca    0.313524   \n",
      "3         Asp79 To Ser319      Asp79     Ser319    0.204270   \n",
      "0         Asn322 To Asp79     Asn322      Asp79    0.179180   \n",
      "6         Leu75 To Ser319      Leu75     Ser319   -0.171330   \n",
      "9  Ser120_Ca To Ser319_Ca  Ser120_Ca  Ser319_Ca   -0.164652   \n",
      "1         Asp79 To Pro323      Asp79     Pro323    0.142609   \n",
      "8  Phe289_Ca To Trp313_Ca  Phe289_Ca  Trp313_Ca   -0.075873   \n",
      "4        Cys116 To Phe282     Cys116     Phe282   -0.070086   \n",
      "\n",
      "                  feature  \n",
      "2         ASP79 to SER120  \n",
      "5         ILE47 to SER319  \n",
      "7   ASN51_CA to GLY320_CA  \n",
      "3         ASP79 to SER319  \n",
      "0         ASN322 to ASP79  \n",
      "6         LEU75 to SER319  \n",
      "9  SER120_CA to SER319_CA  \n",
      "1         ASP79 to PRO323  \n",
      "8  PHE289_CA to TRP313_CA  \n",
      "4        CYS116 to PHE282  \n",
      "Using dark_background\n",
      "             residue  importance\n",
      "Ser120        Ser120    0.408458\n",
      "Asp79          Asp79    0.377829\n",
      "Ile47          Ile47    0.332933\n",
      "Ser319        Ser319    0.320067\n",
      "Gly320_Ca  Gly320_Ca    0.313524\n",
      "Asn51_Ca    Asn51_Ca    0.313524\n",
      "Asn322        Asn322    0.179180\n",
      "Leu75          Leu75   -0.171330\n",
      "Ser319_Ca  Ser319_Ca   -0.164652\n",
      "Ser120_Ca  Ser120_Ca   -0.164652\n",
      "Pro323        Pro323    0.142609\n",
      "Trp313_Ca  Trp313_Ca   -0.075873\n",
      "Phe289_Ca  Phe289_Ca   -0.075873\n",
      "Phe282        Phe282   -0.070086\n",
      "Cys116        Cys116   -0.070086\n",
      "Using dark_background\n",
      "Interpreting tIC 9\n",
      "feature_importances_df.shape\n",
      "(16, 5)\n",
      "residue_importances_df.shape\n",
      "(26, 2)\n",
      "              feature_name      res_i      res_j  importance  \\\n",
      "8     Gly37_Ca To Leu95_Ca   Gly37_Ca   Leu95_Ca    0.444762   \n",
      "11    Ile38_Ca To Leu95_Ca   Ile38_Ca   Leu95_Ca   -0.439249   \n",
      "1           Gly37 To Leu95      Gly37      Leu95    0.394477   \n",
      "2           Ile38 To Leu95      Ile38      Leu95   -0.384963   \n",
      "13  Leu284_Ca To Met279_Ca  Leu284_Ca  Met279_Ca    0.322689   \n",
      "10  Ile278_Ca To Met215_Ca  Ile278_Ca  Met215_Ca    0.306597   \n",
      "9   Ile127_Ca To Tyr326_Ca  Ile127_Ca  Tyr326_Ca    0.208008   \n",
      "14  Met215_Ca To Met279_Ca  Met215_Ca  Met279_Ca   -0.203245   \n",
      "4          Phe89 To Trp105      Phe89     Trp105    0.159163   \n",
      "15   Tyr326_Ca To Val54_Ca  Tyr326_Ca   Val54_Ca   -0.143719   \n",
      "7   Gln170_Ca To Phe104_Ca  Gln170_Ca  Phe104_Ca   -0.135975   \n",
      "3           Leu95 To Val34      Leu95      Val34   -0.132029   \n",
      "12   Ile47_Ca To Tyr316_Ca   Ile47_Ca  Tyr316_Ca   -0.129143   \n",
      "5   Asn312_Ca To Val117_Ca  Asn312_Ca  Val117_Ca    0.108324   \n",
      "0         Arg221 To Thr136     Arg221     Thr136   -0.085688   \n",
      "6   Asn322_Ca To Ile277_Ca  Asn322_Ca  Ile277_Ca    0.078047   \n",
      "\n",
      "                   feature  \n",
      "8     GLY37_CA to LEU95_CA  \n",
      "11    ILE38_CA to LEU95_CA  \n",
      "1           GLY37 to LEU95  \n",
      "2           ILE38 to LEU95  \n",
      "13  LEU284_CA to MET279_CA  \n",
      "10  ILE278_CA to MET215_CA  \n",
      "9   ILE127_CA to TYR326_CA  \n",
      "14  MET215_CA to MET279_CA  \n",
      "4          PHE89 to TRP105  \n",
      "15   TYR326_CA to VAL54_CA  \n",
      "7   GLN170_CA to PHE104_CA  \n",
      "3           LEU95 to VAL34  \n",
      "12   ILE47_CA to TYR316_CA  \n",
      "5   ASN312_CA to VAL117_CA  \n",
      "0         ARG221 to THR136  \n",
      "6   ASN322_CA to ILE277_CA  \n",
      "Using dark_background\n",
      "             residue  importance\n",
      "Gly37_Ca    Gly37_Ca    0.444762\n",
      "Ile38_Ca    Ile38_Ca   -0.439249\n",
      "Leu95_Ca    Leu95_Ca    0.400561\n",
      "Gly37          Gly37    0.394477\n",
      "Ile38          Ile38   -0.384963\n",
      "Leu95          Leu95    0.341827\n",
      "Leu284_Ca  Leu284_Ca    0.322689\n",
      "Ile278_Ca  Ile278_Ca    0.306597\n",
      "Met279_Ca  Met279_Ca    0.296392\n",
      "Met215_Ca  Met215_Ca    0.281105\n",
      "Ile127_Ca  Ile127_Ca    0.208008\n",
      "Tyr326_Ca  Tyr326_Ca    0.190422\n",
      "Trp105        Trp105    0.159163\n",
      "Phe89          Phe89    0.159163\n",
      "Val54_Ca    Val54_Ca   -0.143719\n",
      "Gln170_Ca  Gln170_Ca   -0.135975\n",
      "Phe104_Ca  Phe104_Ca   -0.135975\n",
      "Val34          Val34   -0.132029\n",
      "Ile47_Ca    Ile47_Ca   -0.129143\n",
      "Tyr316_Ca  Tyr316_Ca   -0.129143\n",
      "Val117_Ca  Val117_Ca    0.108324\n",
      "Asn312_Ca  Asn312_Ca    0.108324\n",
      "Thr136        Thr136   -0.085688\n",
      "Arg221        Arg221   -0.085688\n",
      "Asn322_Ca  Asn322_Ca    0.078047\n",
      "Ile277_Ca  Ile277_Ca    0.078047\n",
      "Using dark_background\n",
      "Interpreting tIC 10\n",
      "feature_importances_df.shape\n",
      "(14, 5)\n",
      "residue_importances_df.shape\n",
      "(23, 2)\n",
      "              feature_name      res_i      res_j  importance  \\\n",
      "9     Ile94_Ca To Trp32_Ca   Ile94_Ca   Trp32_Ca    0.369419   \n",
      "10  Leu284_Ca To Met279_Ca  Leu284_Ca  Met279_Ca    0.332701   \n",
      "8   Ile278_Ca To Thr283_Ca  Ile278_Ca  Thr283_Ca    0.322553   \n",
      "13  Phe332_Ca To Tyr326_Ca  Phe332_Ca  Tyr326_Ca    0.293911   \n",
      "2          Ile47 To Ser319      Ile47     Ser319    0.282153   \n",
      "6     Ala91_Ca To Met36_Ca   Ala91_Ca   Met36_Ca   -0.247019   \n",
      "3           Ile94 To Met40      Ile94      Met40   -0.245158   \n",
      "5     Ala91_Ca To Gly37_Ca   Ala91_Ca   Gly37_Ca    0.202282   \n",
      "1          Asn51 To Ser319      Asn51     Ser319   -0.192225   \n",
      "12    Leu95_Ca To Trp32_Ca   Leu95_Ca   Trp32_Ca   -0.166768   \n",
      "0          Ala46 To Phe321      Ala46     Phe321   -0.152207   \n",
      "4           Ile94 To Trp32      Ile94      Trp32   -0.149845   \n",
      "11    Leu95_Ca To Met40_Ca   Leu95_Ca   Met40_Ca   -0.094662   \n",
      "7   Arg131_Ca To Tyr141_Ca  Arg131_Ca  Tyr141_Ca   -0.071821   \n",
      "\n",
      "                   feature  \n",
      "9     ILE94_CA to TRP32_CA  \n",
      "10  LEU284_CA to MET279_CA  \n",
      "8   ILE278_CA to THR283_CA  \n",
      "13  PHE332_CA to TYR326_CA  \n",
      "2          ILE47 to SER319  \n",
      "6     ALA91_CA to MET36_CA  \n",
      "3           ILE94 to MET40  \n",
      "5     ALA91_CA to GLY37_CA  \n",
      "1          ASN51 to SER319  \n",
      "12    LEU95_CA to TRP32_CA  \n",
      "0          ALA46 to PHE321  \n",
      "4           ILE94 to TRP32  \n",
      "11    LEU95_CA to MET40_CA  \n",
      "7   ARG131_CA to TYR141_CA  \n",
      "Using dark_background\n",
      "             residue  importance\n",
      "Ile94_Ca    Ile94_Ca    0.369419\n",
      "Trp32_Ca    Trp32_Ca    0.342609\n",
      "Met279_Ca  Met279_Ca    0.332701\n",
      "Leu284_Ca  Leu284_Ca    0.332701\n",
      "Thr283_Ca  Thr283_Ca    0.322553\n",
      "Ile278_Ca  Ile278_Ca    0.322553\n",
      "Phe332_Ca  Phe332_Ca    0.293911\n",
      "Tyr326_Ca  Tyr326_Ca    0.293911\n",
      "Ile47          Ile47    0.282153\n",
      "Ser319        Ser319    0.258434\n",
      "Met36_Ca    Met36_Ca   -0.247019\n",
      "Met40          Met40   -0.245158\n",
      "Gly37_Ca    Gly37_Ca    0.202282\n",
      "Asn51          Asn51   -0.192225\n",
      "Ala91_Ca    Ala91_Ca    0.179817\n",
      "Ile94          Ile94   -0.154611\n",
      "Phe321        Phe321   -0.152207\n",
      "Ala46          Ala46   -0.152207\n",
      "Trp32          Trp32   -0.149845\n",
      "Leu95_Ca    Leu95_Ca   -0.098268\n",
      "Met40_Ca    Met40_Ca   -0.094662\n",
      "Tyr141_Ca  Tyr141_Ca   -0.071821\n",
      "Arg131_Ca  Arg131_Ca   -0.071821\n",
      "Using dark_background\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9d2227c978>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#import interpret_tICs\n",
    "#reload(interpret_tICs)\n",
    "#from interpret_tICs import *\n",
    "\n",
    "tic_components_dir = tica_dir\n",
    "important_contact_features = interpret_tIC_components(projection_operator_dir, tic_components_dir, feature_residues_pkl, n_tica_components=n_components, percentile=95)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import interpret_tICs\n",
    "reload(interpret_tICs)\n",
    "from interpret_tICs import *\n",
    "tic_subsampled_features_file = \"%s/features_subsampled.pkl\" % tica_dir\n",
    "subsampled_features_dir = os.path.join(tica_dir, \"subsampled_features\")\n",
    "if not os.path.exists(subsampled_features_dir): os.makedirs(subsampled_features_dir)\n",
    "important_contact_features_pruned, important_contact_features_indices = find_non_zero_features(important_contact_features[0], feature_names)\n",
    "if not os.path.exists(tic_subsampled_features_file):\n",
    "    subsample_features(features_dir, important_contact_features_indices, important_contact_features_pruned, tic_subsampled_features_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/features_subsampled.pkl'"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tic_subsampled_features_file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loading /home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/features_subsampled.pkl\n"
     ]
    }
   ],
   "source": [
    "protein_top_features = load_file(tic_subsampled_features_file)\n",
    "protein_top_features_concat = pd.concat(protein_top_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "salm_trajs = []\n",
    "apo_trajs = []\n",
    "salm_nb_trajs = []\n",
    "trajs = get_trajectory_files(traj_dir, \".h5\")\n",
    "#trajs = [t for t in get_trajectory_files(traj_dir, traj_ext) if \"apo_reimaged\" not in t and \"salm_nb\" not in t]\n",
    "for i, traj in enumerate(trajs):\n",
    "    if \"NNNS\" in traj or \"APO\" in traj:\n",
    "        apo_trajs.append(i)\n",
    "    elif \"NNWS\" in traj or \"nonbWL\" in traj:\n",
    "        salm_trajs.append(i)\n",
    "    elif \"WNWS\" in traj or \"RWNW\" in traj:\n",
    "        salm_nb_trajs.append(i)\n",
    "\n",
    "#user_defined_coords = load_file(user_defined_features_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#salm_pnas = [user_defined_coords[i] for i in salm_trajs]\n",
    "#apo_pnas = [user_defined_coords[i] for i in apo_trajs]\n",
    "#salm_nb_pnas = [user_defined_coords[i] for i in salm_nb_trajs]\n",
    "#salm_nb_pnas = [user_defined_coords[i] for i in salm_nb_trajs]\n",
    "#f338a_pnas = [user_defined_coords[i] for i in f338a_trajs]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#salm_pnas_file = \"%s/salm.h5\" % whole_trajectory_pnas\n",
    "#apo_pnas_file = \"%s/apo.h5\" % whole_trajectory_pnas\n",
    "#salm_nb_pnas_file = \"%s/salm_nb.h5\" % whole_trajectory_pnas\n",
    "#salm_nb_pnas_file = \"%s/salm_nb.h5\" % whole_trajectory_pnas\n",
    "#f338a_pnas_file = \"%s/f338a.h5\" % whole_trajectory_pnas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'\\nfrom imp import reload\\nimport analysis\\nreload(analysis)\\nfrom analysis import *\\nimport seaborn as sns\\nsalm_pnas_concatenated = np.concatenate(salm_pnas)\\nprint(np.shape(salm_pnas_concatenated))\\napo_pnas_concatenated = np.concatenate(apo_pnas)\\n#salm_nb_pnas_concatenated = np.concatenate(salm_nb_pnas)\\nsuf_pnas_concatenated = np.concatenate(suf_pnas)\\nf338a_pnas_concatenated = np.concatenate(f338a_pnas)\\n\\nverbosedump(salm_pnas, salm_pnas_file)\\nverbosedump(apo_pnas, apo_pnas_file)\\n#verbosedump(salm_nb_pnas, salm_nb_pnas_file)\\nverbosedump(suf_pnas, suf_pnas_file)\\nverbosedump(f338a_pnas, f338a_pnas_file)\\n\\n\\ncolumns=sorted(feature_name_residues_dict.keys())\\n#plot_columns(whole_trajectory_pnas, salm_pnas_file, titles=columns, main = \"salm\", tICA = False, scale = 1.0, refcoords_file = None)\\n#plot_columns(whole_trajectory_pnas, apo_pnas_file, titles=columns, main = \"Apo\", tICA = False, scale = 1.0, refcoords_file = None)\\n\\n#df = pd.DataFrame(salm_pnas_concatenated, columns=sorted(feature_name_residues_dict.keys()))\\n#sns.jointplot(x=\"tm6_tm3_dist\", y=\"rmsd_npxxy_inactive\", data=df, kind=\"kde\")\\n#plt.hexbin(salm_pnas_concatenated[:,0], salm_pnas_concatenated[:,2], cmap=plt.cm.YlOrRd_r)\\n'"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "from imp import reload\n",
    "import analysis\n",
    "reload(analysis)\n",
    "from analysis import *\n",
    "import seaborn as sns\n",
    "salm_pnas_concatenated = np.concatenate(salm_pnas)\n",
    "print(np.shape(salm_pnas_concatenated))\n",
    "apo_pnas_concatenated = np.concatenate(apo_pnas)\n",
    "#salm_nb_pnas_concatenated = np.concatenate(salm_nb_pnas)\n",
    "salm_nb_pnas_concatenated = np.concatenate(salm_nb_pnas)\n",
    "f338a_pnas_concatenated = np.concatenate(f338a_pnas)\n",
    "\n",
    "verbosedump(salm_pnas, salm_pnas_file)\n",
    "verbosedump(apo_pnas, apo_pnas_file)\n",
    "#verbosedump(salm_nb_pnas, salm_nb_pnas_file)\n",
    "verbosedump(salm_nb_pnas, salm_nb_pnas_file)\n",
    "verbosedump(f338a_pnas, f338a_pnas_file)\n",
    "\n",
    "\n",
    "columns=sorted(feature_name_residues_dict.keys())\n",
    "#plot_columns(whole_trajectory_pnas, salm_pnas_file, titles=columns, main = \"salm\", tICA = False, scale = 1.0, refcoords_file = None)\n",
    "#plot_columns(whole_trajectory_pnas, apo_pnas_file, titles=columns, main = \"Apo\", tICA = False, scale = 1.0, refcoords_file = None)\n",
    "\n",
    "#df = pd.DataFrame(salm_pnas_concatenated, columns=sorted(feature_name_residues_dict.keys()))\n",
    "#sns.jointplot(x=\"tm6_tm3_dist\", y=\"rmsd_npxxy_inactive\", data=df, kind=\"kde\")\n",
    "#plt.hexbin(salm_pnas_concatenated[:,0], salm_pnas_concatenated[:,2], cmap=plt.cm.YlOrRd_r)\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "salm_nb_tica_file = \"%s/salm_nb_tica.h5\" % tica_dir\n",
    "salm_tica_file = \"%s/salm_tica.h5\" % tica_dir\n",
    "apo_tica_file = \"%s/apo_tica.h5\" % tica_dir\n",
    "#salm_nb_tica_file = \"%s/salm_nb_tica.h5\" % tica_dir"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/salm_nb_tica.h5\"... (<class 'list'>)\n",
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/salm_tica.h5\"... (<class 'list'>)\n",
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/apo_tica.h5\"... (<class 'list'>)\n"
     ]
    }
   ],
   "source": [
    "tica_coords = compat_verboseload(projected_features_dir)\n",
    "salm_nb_tica = [tica_coords[i] for i in salm_nb_trajs]\n",
    "salm_tica = [tica_coords[i] for i in salm_trajs]\n",
    "apo_tica = [tica_coords[i] for i in apo_trajs]\n",
    "\n",
    "\n",
    "verbosedump(salm_nb_tica, salm_nb_tica_file)\n",
    "verbosedump(salm_tica, salm_tica_file)\n",
    "verbosedump(apo_tica, apo_tica_file)\n",
    "\n",
    "columns = [\"tIC%d\" %i for i in range(1,n_components+1)]\n",
    "#plot_columns(tica_dir, salm_tica_file, titles=columns, main = \"salm\", tICA = False, scale = 1.0, refcoords_file = None)\n",
    "#plot_columns(tica_dir, apo_tica_file, titles=columns, main = \"Apo\", tICA = False, scale = 1.0, refcoords_file = None)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(47939, 10)\n",
      "(50340, 10)\n",
      "(41653, 10)\n"
     ]
    }
   ],
   "source": [
    "print(np.concatenate(salm_nb_tica).shape)\n",
    "print(np.concatenate(apo_tica).shape)\n",
    "print(np.concatenate(salm_tica).shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import custom_clusterer\n",
    "reload(custom_clusterer)\n",
    "from custom_clusterer import *\n",
    "n_clusters = 100\n",
    "clusterer_dir = \"%s/clusterer_%dclusters.h5\" %(tica_dir, n_clusters)\n",
    "if not os.path.exists(clusterer_dir):\n",
    "    cluster_minikmeans(tica_dir, projected_features_dir, traj_dir, n_clusters=n_clusters, clusterer_dir=clusterer_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/clusterer_100clusters.h5\n"
     ]
    }
   ],
   "source": [
    "print(clusterer_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "salm_nb_clusterer_file = \"%s/salm_nb_clusterer%d.h5\" %(tica_dir, n_clusters)\n",
    "salm_clusterer_file = \"%s/salm_clusterer%d.h5\" %(tica_dir, n_clusters)\n",
    "apo_clusterer_file = \"%s/apo_clusterer%d.h5\" %(tica_dir, n_clusters)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/salm_clusterer100.h5\"... (<class 'msmbuilder.cluster.MiniBatchKMeans'>)\n",
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/apo_clusterer100.h5\"... (<class 'msmbuilder.cluster.MiniBatchKMeans'>)\n",
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/salm_nb_clusterer100.h5\"... (<class 'msmbuilder.cluster.MiniBatchKMeans'>)\n"
     ]
    }
   ],
   "source": [
    "clusterer = compat_verboseload(clusterer_dir)\n",
    "cluster_labels = clusterer.labels_\n",
    "salm_nb_clusters = [cluster_labels[i] for i in salm_nb_trajs]\n",
    "salm_clusters = [cluster_labels[i] for i in salm_trajs]\n",
    "apo_clusters = [cluster_labels[i] for i in apo_trajs]\n",
    "\n",
    "salm_clusterer = copy.deepcopy(clusterer)\n",
    "salm_clusterer.labels_ = salm_clusters\n",
    "#if not os.path.exists(salm_clusterer_file):\n",
    "verbosedump(salm_clusterer, salm_clusterer_file)\n",
    "\n",
    "apo_clusterer = copy.deepcopy(clusterer)\n",
    "apo_clusterer.labels_ = apo_clusters\n",
    "    \n",
    "#if not os.path.exists(apo_clusterer_file):\n",
    "verbosedump(apo_clusterer, apo_clusterer_file)\n",
    "\n",
    "salm_nb_clusterer = copy.deepcopy(clusterer)\n",
    "salm_nb_clusterer.labels_ = salm_nb_clusters\n",
    "    \n",
    "#if not os.path.exists(salm_nb_clusterer_file):\n",
    "verbosedump(salm_nb_clusterer, salm_nb_clusterer_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "cluster_tica_averages = calculate_cluster_averages_per_feature(salm_clusterer, salm_tica)\n",
    "cluster_tica_averages = pd.DataFrame(cluster_tica_averages, columns=[\"tIC.%d\" %i for i in range(1, n_components+1)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import custom_msm\n",
    "reload(custom_msm)\n",
    "from custom_msm import *\n",
    "from msm_resampled import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1010\n",
      "1144\n",
      "251\n",
      "562\n",
      "418\n",
      "344\n",
      "1111\n",
      "1183\n",
      "0\n",
      "240\n",
      "700\n",
      "22\n",
      "0\n",
      "129\n",
      "221\n",
      "0\n",
      "460\n",
      "232\n",
      "86\n",
      "1043\n",
      "1075\n",
      "195\n",
      "2118\n",
      "851\n",
      "240\n",
      "853\n",
      "25\n",
      "879\n",
      "577\n",
      "4\n",
      "9\n",
      "25\n",
      "326\n",
      "814\n",
      "113\n",
      "985\n",
      "107\n",
      "25\n",
      "212\n",
      "107\n",
      "1158\n",
      "119\n",
      "330\n",
      "21\n",
      "74\n",
      "726\n",
      "1110\n",
      "635\n",
      "1165\n",
      "17\n",
      "400\n",
      "509\n",
      "69\n",
      "64\n",
      "903\n",
      "240\n",
      "749\n",
      "807\n",
      "1228\n",
      "196\n",
      "354\n",
      "246\n",
      "8\n",
      "333\n",
      "110\n",
      "578\n",
      "414\n",
      "1075\n",
      "292\n",
      "851\n",
      "805\n",
      "476\n",
      "590\n",
      "322\n",
      "81\n",
      "10\n",
      "209\n",
      "194\n",
      "356\n",
      "993\n",
      "757\n",
      "5\n",
      "1238\n",
      "1024\n",
      "254\n",
      "503\n",
      "439\n",
      "1035\n",
      "629\n",
      "587\n",
      "1\n",
      "1148\n",
      "522\n",
      "465\n",
      "387\n",
      "125\n",
      "932\n",
      "2068\n",
      "276\n",
      "432\n",
      "61\n",
      "1089\n",
      "0\n",
      "504\n",
      "822\n",
      "318\n",
      "0\n",
      "266\n",
      "879\n",
      "0\n",
      "308\n",
      "1348\n",
      "895\n",
      "64\n",
      "6\n",
      "556\n",
      "39\n",
      "1160\n",
      "1\n",
      "473\n",
      "0\n",
      "198\n",
      "737\n",
      "104\n",
      "831\n",
      "915\n",
      "1739\n",
      "22\n",
      "294\n",
      "105\n",
      "661\n",
      "724\n",
      "0\n",
      "1380\n",
      "245\n",
      "498\n",
      "61\n",
      "153\n",
      "279\n",
      "671\n",
      "1534\n",
      "16\n",
      "3\n",
      "711\n",
      "10\n",
      "83\n",
      "30\n",
      "229\n",
      "1813\n",
      "106\n",
      "102\n",
      "1388\n",
      "0\n",
      "0\n",
      "437\n",
      "0\n",
      "75\n",
      "244\n",
      "237\n",
      "335\n",
      "67\n",
      "195\n",
      "300\n",
      "56\n",
      "51\n",
      "221\n",
      "296\n",
      "2\n",
      "162\n",
      "553\n",
      "810\n",
      "4\n",
      "1277\n",
      "8\n",
      "523\n",
      "39\n",
      "0\n",
      "241\n",
      "366\n",
      "1047\n",
      "293\n",
      "15\n",
      "1785\n",
      "1774\n",
      "292\n",
      "69\n",
      "0\n",
      "909\n",
      "531\n",
      "57\n",
      "7\n",
      "299\n",
      "681\n",
      "0\n",
      "206\n",
      "307\n",
      "1224\n",
      "549\n",
      "678\n",
      "0\n",
      "46\n",
      "1192\n",
      "0\n",
      "6930\n",
      "102\n",
      "629\n",
      "0\n",
      "1433\n",
      "3\n",
      "0\n",
      "33\n",
      "776\n",
      "57\n",
      "2159\n",
      "193\n",
      "619\n",
      "124\n",
      "0\n",
      "0\n",
      "1614\n",
      "0\n",
      "87\n",
      "69\n",
      "0\n",
      "0\n",
      "253\n",
      "0\n",
      "6\n",
      "72\n",
      "18\n",
      "0\n",
      "11\n",
      "0\n",
      "51\n",
      "1825\n",
      "149\n",
      "1066\n",
      "248\n",
      "1362\n",
      "229\n",
      "278\n",
      "201\n",
      "0\n",
      "233\n",
      "632\n",
      "240\n",
      "1\n",
      "0\n",
      "996\n",
      "0\n",
      "59\n",
      "2353\n",
      "0\n",
      "304\n",
      "756\n",
      "86\n",
      "1347\n",
      "173\n",
      "543\n",
      "756\n",
      "192\n",
      "58\n",
      "0\n",
      "2028\n",
      "9\n",
      "19\n",
      "0\n",
      "0\n",
      "1054\n",
      "535\n",
      "245\n",
      "37\n",
      "623\n",
      "0\n",
      "708\n",
      "342\n",
      "0\n",
      "279\n",
      "1060\n",
      "415\n",
      "292\n",
      "395\n",
      "3212\n",
      "107\n",
      "428\n",
      "45\n",
      "0\n",
      "1142\n",
      "31\n",
      "3\n",
      "198\n",
      "30\n",
      "108\n",
      "0\n",
      "1128\n",
      "476\n",
      "1438\n",
      "528\n",
      "460\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "apo_clusters_map = make_clusters_map(apo_clusterer)\n",
    "salm_clusters_map = make_clusters_map(salm_clusterer)\n",
    "salm_nb_clusters_map = make_clusters_map(salm_nb_clusterer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#plt.clf()\n",
    "#plot_timescales(apo_clusterer_file, n_clusters, tica_dir, main=\"Apo MSM Timescales\", lag_times=list(range(1,51,10)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#plt.clf()\n",
    "#import custom_msm\n",
    "#reload(custom_msm)\n",
    "#from custom_msm import *\n",
    "#plot_timescales(salm_clusterer_file, n_clusters, tica_dir, main=\"salm MSM Timescales\", lag_times=[1,5,10,15,20])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#plt.clf()\n",
    "#plot_timescales(salm_nb_clusterer_file, n_clusters, tica_dir, main=\"salm_nb MSM Timescales\", lag_times=list(range(1,51,10)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loading \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/apo_clusterer100.h5\"...\n",
      "fitting msm to trajectories with 100 clusters and lag_time 25\n",
      "MarkovStateModel(ergodic_cutoff='off', lag_time=25.0, n_timescales=None,\n",
      "         prior_counts=1e-05, reversible_type='mle', sliding_window=True,\n",
      "         verbose=True)\n",
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/apo_msm_lag-time25_prior-counts1e-05_clusters100.h5\"... (<class 'msmbuilder.msm.msm.MarkovStateModel'>)\n",
      "fitted msm to trajectories with 97 states\n",
      "Found number to sample per cluster based on equilibrium proporrtions.\n",
      "Obtained random (trajectory, frame) pairs based on equilibrium populations\n",
      "Rearranged equilibrium sampled frames based on trajectories\n",
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/apo_msm_lag-time25_prior-counts1e-05_clusters100_resampled_20000.h5\"... (<class 'dict'>)\n",
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/apo_tica_msm_lag-time25_prior-counts1e-05_clusters100_resampled_20000.h5\"... (<class 'numpy.ndarray'>)\n"
     ]
    }
   ],
   "source": [
    "import custom_msm\n",
    "import importlib\n",
    "importlib.reload(custom_msm)\n",
    "from custom_msm import *\n",
    "\n",
    "apo_lag_time=25.\n",
    "prior_counts = 1e-5\n",
    "ergodic_cutoff = 'off'\n",
    "apo_msm_dir = os.path.join(tica_dir, \"apo_msm_lag-time%d_prior-counts%s_clusters%d.h5\" %(apo_lag_time, str(prior_counts), n_clusters))\n",
    "#if 1==1:\n",
    "if not os.path.exists(apo_msm_dir):\n",
    "    apo_msm = build_msm(apo_clusterer_file, apo_lag_time, apo_msm_dir, prior_counts, ergodic_cutoff)\n",
    "else:\n",
    "    apo_msm = compat_verboseload(apo_msm_dir)\n",
    "\n",
    "import msm_resampled\n",
    "reload(msm_resampled)\n",
    "from msm_resampled import *\n",
    "total_samples = 20000\n",
    "num_trajs = len(apo_trajs)\n",
    "apo_resampled_traj_to_frames_file = os.path.join(tica_dir, \"apo_msm_lag-time%d_prior-counts%s_clusters%d_resampled_%d.h5\" %(apo_lag_time, str(prior_counts), n_clusters, total_samples))\n",
    "#if not os.path.exists(apo_resampled_traj_to_frames_file):\n",
    "apo_resampled_traj_to_frames = resample_by_msm(total_samples, apo_msm, apo_clusters_map, num_trajs, apo_resampled_traj_to_frames_file)\n",
    "#else:\n",
    "apo_resampled_traj_to_frames = compat_verboseload(apo_resampled_traj_to_frames_file)\n",
    "\n",
    "#apo_pnas = compat_verboseload(apo_pnas_file)\n",
    "#apo_pnas_resampled_file = os.path.join(tica_dir, \"apo_pnas_msm_lag-time%d_prior-counts%s_clusters%d_resampled_%d.h5\" %(apo_lag_time, str(prior_counts), n_clusters, total_samples))\n",
    "#resample_features_by_msm_equilibirum_pop(apo_pnas, apo_resampled_traj_to_frames, apo_pnas_resampled_file)\n",
    "#apo_pnas_resampled = compat_verboseload(apo_pnas_resampled_file)\n",
    "\n",
    "tica_concatenated = np.concatenate(compat_verboseload(projected_features_dir))\n",
    "tica_axes = [(np.min(tica_concatenated[:,i]), np.max(tica_concatenated[:,i])) for i in range(0, np.shape(tica_concatenated)[1])]\n",
    "\n",
    "apo_tica = compat_verboseload(apo_tica_file)\n",
    "apo_tica_resampled_file = os.path.join(tica_dir, \"apo_tica_msm_lag-time%d_prior-counts%s_clusters%d_resampled_%d.h5\" %(apo_lag_time, str(prior_counts), n_clusters, total_samples))\n",
    "resample_features_by_msm_equilibirum_pop(apo_tica, apo_resampled_traj_to_frames, apo_tica_resampled_file)\n",
    "apo_tica_resampled = compat_verboseload(apo_tica_resampled_file)\n",
    "\n",
    "apo_clusterer = compat_verboseload(apo_clusterer_file)\n",
    "apo_msm = compat_verboseload(apo_msm_dir)\n",
    "\n",
    "cluster_id_proportion_tuples = []\n",
    "\n",
    "equilibrium_populations = apo_msm.populations_\n",
    "\n",
    "for cluster_id in apo_msm.mapping_.keys():\n",
    "    state_id = apo_msm.mapping_[cluster_id]\n",
    "    cluster_id_proportion_tuples.append((cluster_id,equilibrium_populations[state_id]))\n",
    "cluster_id_proportion_tuples = sorted(cluster_id_proportion_tuples, key=itemgetter(1), reverse=True)\n",
    "\n",
    "#plot_all_tics_and_clusters(tica_dir, apo_tica_resampled_file, apo_clusterer_file, lag_time, tic_range=range(0,n_components), main = \"Apo_MSM_Reweighted\", label = \"cluster_id\", active_cluster_ids = [c[0] for c in cluster_id_proportion_tuples[0:100]], intermediate_cluster_ids = [], inactive_cluster_ids = [], concatenate=False, axes=tica_axes)\n",
    "\n",
    "#plot_columns(tica_dir, apo_tica_resampled_file, titles = [\"tIC.%d\" %i for i in range(1,n_components+1)], main = \"mOR_Apo_MSM_\", tICA = False, scale = 1.0, refcoords_file = ref_tica_coords, axes=tica_axes, concatenate=False, reshape=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loading \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/salm_clusterer100.h5\"...\n",
      "fitting msm to trajectories with 100 clusters and lag_time 25\n",
      "MarkovStateModel(ergodic_cutoff='off', lag_time=25, n_timescales=None,\n",
      "         prior_counts=1e-05, reversible_type='mle', sliding_window=True,\n",
      "         verbose=True)\n",
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/salm_msm_lag-time25_prior-counts1e-05_clusters100.h5\"... (<class 'msmbuilder.msm.msm.MarkovStateModel'>)\n",
      "fitted msm to trajectories with 88 states\n",
      "Found number to sample per cluster based on equilibrium proporrtions.\n",
      "Obtained random (trajectory, frame) pairs based on equilibrium populations\n",
      "Rearranged equilibrium sampled frames based on trajectories\n",
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/salm_msm_lag-time25_prior-counts1e-05_clusters100_resampled_20000.h5\"... (<class 'dict'>)\n",
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/salm_tica_msm_lag-time25_prior-counts1e-05_clusters100_resampled_20000.h5\"... (<class 'numpy.ndarray'>)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "salm_lag_time=25\n",
    "prior_counts = 1e-5\n",
    "ergodic_cutoff='off'\n",
    "salm_msm_dir = os.path.join(tica_dir, \"salm_msm_lag-time%d_prior-counts%s_clusters%d.h5\" %(salm_lag_time, str(prior_counts), n_clusters))\n",
    "#if 1==1:\n",
    "if not os.path.exists(salm_msm_dir):\n",
    "    salm_msm = build_msm(salm_clusterer_file, salm_lag_time, salm_msm_dir, prior_counts, ergodic_cutoff)\n",
    "else:\n",
    "    salm_msm = compat_verboseload(salm_msm_dir)\n",
    "\n",
    "import msm_resampled\n",
    "reload(msm_resampled)\n",
    "from msm_resampled import *\n",
    "total_samples = 20000\n",
    "num_trajs = len(salm_trajs)\n",
    "salm_resampled_traj_to_frames_file = os.path.join(tica_dir, \"salm_msm_lag-time%d_prior-counts%s_clusters%d_resampled_%d.h5\" %(salm_lag_time, str(prior_counts), n_clusters, total_samples))\n",
    "#if not os.path.exists(salm_resampled_traj_to_frames_file):\n",
    "salm_resampled_traj_to_frames = resample_by_msm(total_samples, salm_msm, salm_clusters_map, num_trajs, salm_resampled_traj_to_frames_file)\n",
    "#else:\n",
    "salm_resampled_traj_to_frames = compat_verboseload(salm_resampled_traj_to_frames_file)\n",
    "\n",
    "#salm_pnas = compat_verboseload(salm_pnas_file)\n",
    "#salm_pnas_resampled_file = os.path.join(tica_dir, \"salm_pnas_msm_lag-time%d_prior-counts%s_clusters%d_resampled_%d.h5\" %(salm_lag_time, str(prior_counts), n_clusters, total_samples))\n",
    "#resample_features_by_msm_equilibirum_pop(salm_pnas, salm_resampled_traj_to_frames, salm_pnas_resampled_file)\n",
    "#salm_pnas_resampled = compat_verboseload(salm_pnas_resampled_file)\n",
    "\n",
    "tica_concatenated = np.concatenate(compat_verboseload(projected_features_dir))\n",
    "tica_axes = [(np.min(tica_concatenated[:,i]), np.max(tica_concatenated[:,i])) for i in range(0, np.shape(tica_concatenated)[1])]\n",
    "\n",
    "salm_tica = compat_verboseload(salm_tica_file)\n",
    "salm_tica_resampled_file = os.path.join(tica_dir, \"salm_tica_msm_lag-time%d_prior-counts%s_clusters%d_resampled_%d.h5\" %(salm_lag_time, str(prior_counts), n_clusters, total_samples))\n",
    "resample_features_by_msm_equilibirum_pop(salm_tica, salm_resampled_traj_to_frames, salm_tica_resampled_file)\n",
    "salm_tica_resampled = compat_verboseload(salm_tica_resampled_file)\n",
    "\n",
    "salm_clusterer = compat_verboseload(salm_clusterer_file)\n",
    "salm_msm = compat_verboseload(salm_msm_dir)\n",
    "\n",
    "cluster_id_proportion_tuples = []\n",
    "\n",
    "equilibrium_populations = salm_msm.populations_\n",
    "\n",
    "for cluster_id in salm_msm.mapping_.keys():\n",
    "    state_id = salm_msm.mapping_[cluster_id]\n",
    "    cluster_id_proportion_tuples.append((cluster_id,equilibrium_populations[state_id]))\n",
    "cluster_id_proportion_tuples = sorted(cluster_id_proportion_tuples, key=itemgetter(1), reverse=True)\n",
    "    \n",
    "#plot_all_tics_and_clusters(tica_dir, salm_tica_resampled_file, salm_clusterer_file, lag_time, tic_range=range(0,n_components), main = \"salm_MSM_Reweighted\", label = \"cluster_id\", active_cluster_ids = [c[0] for c in cluster_id_proportion_tuples[0:100]], intermediate_cluster_ids = [], inactive_cluster_ids = [], concatenate=False, axes=tica_axes)\n",
    "\n",
    "#plot_columns(tica_dir, salm_tica_resampled_file, titles = [\"tIC.%d\" %i for i in range(1,n_components+1)], main = \"mOR_salm_MSM_\", tICA = False, scale = 1.0, refcoords_file = ref_tica_coords, axes=tica_axes, concatenate=False, reshape=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "graph_file = \"%s/salm_clusters_%d_msm_lag_time%d_graph.graphml\" % (tica_dir, n_clusters, salm_lag_time)\n",
    "salm_tica_averages = calculate_cluster_averages_per_feature(salm_clusterer, salm_tica)\n",
    "salm_cluster_attributes = {}\n",
    "for i, title in enumerate([\"tIC.%d\" %i for i in range(1,n_components+1)]):\n",
    "    salm_cluster_attributes[title] = salm_tica_averages[:,i]\n",
    "\n",
    "salm_msm_attributes = {'eq_pops' : salm_msm.populations_}\n",
    "construct_graph(salm_msm_dir, salm_clusterer_file, n_clusters, 5, 5, graph_file, inactive = None, active = None, pnas_clusters_averages = None, tica_clusters_averages = None, docking=None, macrostate = None, cluster_attributes=salm_cluster_attributes, msm_attributes=salm_msm_attributes)\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loading \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/salm_nb_clusterer100.h5\"...\n",
      "fitting msm to trajectories with 100 clusters and lag_time 25\n",
      "MarkovStateModel(ergodic_cutoff='off', lag_time=25, n_timescales=None,\n",
      "         prior_counts=1e-05, reversible_type='mle', sliding_window=True,\n",
      "         verbose=True)\n",
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/salm_nb_msm_lag-time25_prior-counts1e-05_clusters100.h5\"... (<class 'msmbuilder.msm.msm.MarkovStateModel'>)\n",
      "fitted msm to trajectories with 77 states\n",
      "Found number to sample per cluster based on equilibrium proporrtions.\n",
      "Obtained random (trajectory, frame) pairs based on equilibrium populations\n",
      "Rearranged equilibrium sampled frames based on trajectories\n",
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/salm_nb_msm_lag-time25_prior-counts1e-05_clusters100_resampled_20000.h5\"... (<class 'dict'>)\n",
      "Saving \"/home/enf/md_simulations/b2ar/Salmeterol/sparse-tICA_t50_n_components10contact_6pt6_angstroms-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/salm_nb_tica_msm_lag-time25_prior-counts1e-05_clusters100_resampled_20000.h5\"... (<class 'numpy.ndarray'>)\n"
     ]
    }
   ],
   "source": [
    "salm_nb_lag_time=25\n",
    "prior_counts = 1e-5\n",
    "ergodic_cutoff='off'\n",
    "salm_nb_msm_dir = os.path.join(tica_dir, \"salm_nb_msm_lag-time%d_prior-counts%s_clusters%d.h5\" %(salm_nb_lag_time, str(prior_counts), n_clusters))\n",
    "if not os.path.exists(salm_nb_msm_dir):\n",
    "    salm_nb_msm = build_msm(salm_nb_clusterer_file, salm_nb_lag_time, salm_nb_msm_dir, prior_counts, ergodic_cutoff)\n",
    "else:\n",
    "    salm_nb_msm = compat_verboseload(salm_nb_msm_dir)\n",
    "\n",
    "import msm_resampled\n",
    "reload(msm_resampled)\n",
    "from msm_resampled import *\n",
    "total_samples = 20000\n",
    "num_trajs = len(salm_nb_trajs)\n",
    "salm_nb_resampled_traj_to_frames_file = os.path.join(tica_dir, \"salm_nb_msm_lag-time%d_prior-counts%s_clusters%d_resampled_%d.h5\" %(salm_nb_lag_time, str(prior_counts), n_clusters, total_samples))\n",
    "#if not os.path.exists(salm_nb_resampled_traj_to_frames_file):\n",
    "salm_nb_resampled_traj_to_frames = resample_by_msm(total_samples, salm_nb_msm, salm_nb_clusters_map, num_trajs, salm_nb_resampled_traj_to_frames_file)\n",
    "#else:\n",
    "salm_nb_resampled_traj_to_frames = compat_verboseload(salm_nb_resampled_traj_to_frames_file)\n",
    "\n",
    "#salm_nb_pnas = compat_verboseload(salm_nb_pnas_file)\n",
    "#salm_nb_pnas_resampled_file = os.path.join(tica_dir, \"salm_nb_pnas_msm_lag-time%d_prior-counts%s_clusters%d_resampled_%d.h5\" %(salm_nb_lag_time, str(prior_counts), n_clusters, total_samples))\n",
    "#resample_features_by_msm_equilibirum_pop(salm_nb_pnas, salm_nb_resampled_traj_to_frames, salm_nb_pnas_resampled_file)\n",
    "#salm_nb_pnas_resampled = compat_verboseload(salm_nb_pnas_resampled_file)\n",
    "\n",
    "tica_concatenated = np.concatenate(compat_verboseload(projected_features_dir))\n",
    "tica_axes = [(np.min(tica_concatenated[:,i]), np.max(tica_concatenated[:,i])) for i in range(0, np.shape(tica_concatenated)[1])]\n",
    "\n",
    "salm_nb_tica = compat_verboseload(salm_nb_tica_file)\n",
    "salm_nb_tica_resampled_file = os.path.join(tica_dir, \"salm_nb_tica_msm_lag-time%d_prior-counts%s_clusters%d_resampled_%d.h5\" %(salm_nb_lag_time, str(prior_counts), n_clusters, total_samples))\n",
    "resample_features_by_msm_equilibirum_pop(salm_nb_tica, salm_nb_resampled_traj_to_frames, salm_nb_tica_resampled_file)\n",
    "salm_nb_tica_resampled = compat_verboseload(salm_nb_tica_resampled_file)\n",
    "\n",
    "salm_nb_clusterer = compat_verboseload(salm_nb_clusterer_file)\n",
    "salm_nb_msm = compat_verboseload(salm_nb_msm_dir)\n",
    "\n",
    "cluster_id_proportion_tuples = []\n",
    "\n",
    "equilibrium_populations = salm_nb_msm.populations_\n",
    "\n",
    "for cluster_id in salm_nb_msm.mapping_.keys():\n",
    "    state_id = salm_nb_msm.mapping_[cluster_id]\n",
    "    cluster_id_proportion_tuples.append((cluster_id,equilibrium_populations[state_id]))\n",
    "cluster_id_proportion_tuples = sorted(cluster_id_proportion_tuples, key=itemgetter(1), reverse=True)\n",
    "    \n",
    "#plot_all_tics_and_clusters(tica_dir, salm_nb_tica_resampled_file, salm_nb_clusterer_file, lag_time, tic_range=range(0,n_components), main = \"salm_nb_MSM_Reweighted\", label = \"cluster_id\", active_cluster_ids = [c[0] for c in cluster_id_proportion_tuples[0:100]], intermediate_cluster_ids = [], inactive_cluster_ids = [], concatenate=False, axes=tica_axes)\n",
    "\n",
    "#plot_columns(tica_dir, salm_nb_tica_resampled_file, titles = [\"tIC.%d\" %i for i in range(1,n_components+1)], main = \"mOR_salm_nb_MSM_\", tICA = False, scale = 1.0, refcoords_file = ref_tica_coords, axes=tica_axes, concatenate=False, reshape=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "tica_resampled = np.vstack([compat_verboseload(salm_tica_resampled_file)[::1,], compat_verboseload(apo_tica_resampled_file)[::1,], compat_verboseload(salm_nb_tica_resampled_file)[::1,]])\n",
    "mins = np.min(tica_resampled, axis=0)\n",
    "maxs = np.max(tica_resampled, axis=0)\n",
    "stds = np.std(tica_resampled, axis=0)\n",
    "custom_lims = [[mins[i] - .5*stds[i], maxs[i] + .5*stds[i]] for i in range(0,len(mins))]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#ref_pnas = load_file(\"/home/enf/md_simulations/MOR/h8_reimaged/all_pnas_features/ref_coords.h5\")\n",
    "#ref_pnas = np.hstack([np.array(x).T for x in ref_pnas])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loading /home/enf/md_simulations/b2ar/Salmeterol/featurescontact_6pt6_angstroms-CA-py3-far/reference_receptors/2RH1_prepped-A.dataset\n",
      "loading /home/enf/md_simulations/b2ar/Salmeterol/featurescontact_6pt6_angstroms-CA-py3-far/reference_receptors/3P0G_pymol_prepped-A.dataset\n",
      "loading /home/enf/md_simulations/b2ar/Salmeterol/featurescontact_6pt6_angstroms-CA-py3-far/reference_receptors/ionized.dataset\n",
      "(3, 2610)\n"
     ]
    }
   ],
   "source": [
    "def get_features(features_dir):\n",
    "    feature_files = get_trajectory_files(features_dir, \".dataset\")\n",
    "    features = np.vstack([load_file(f) for f in feature_files])\n",
    "    print(features.shape)\n",
    "    return features\n",
    "\n",
    "ref_pp_features = pd.DataFrame(get_features(ref_features_dir), columns=feature_names)\n",
    "\n",
    "ref_pp_tica = pd.read_csv(ref_tica_coords, header=None).values\n",
    "#ref_pp_tica.columns = [\"mOR tIC.%d\" %i for i in range(1,ref_pp_tica.shape[1])]\n",
    "\n",
    "#ref_pnas = load_file(\"/home/enf/md_simulations/MOR/h8_reimaged/all_pnas_features/ref_coords.h5\")\n",
    "#ref_pnas = np.hstack([np.array(x).T for x in ref_pnas])\n",
    "\n",
    "#ref_pnas = pd.DataFrame(ref_pnas, columns=sorted(feature_name_residues_dict.keys()))\n",
    "\n",
    "#ref_features = pd.concat([ref_pp_features*10., ref_lp_features*10., ref_pp_tica, ref_lp_tica, ref_pnas], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#coords_bounds_dict = {\"tIC.4\": [(8.0, 10.), (12., 15.)]}\n",
    "#all_traj_files = get_trajectory_files(traj_dir, traj_ext)\n",
    "#trajs_to_watch = reaction_coordinate_sampler([all_traj_files[i] for i in salm_trajs], \".h5\", [t.values for t in all_traj_features], \n",
    "#                            all_traj_features[0].columns.values.tolist(), coords_bounds_dict, \n",
    "#                            \"%s/trajs_to_watch.csv\" %pp_tica_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Making delta G plots.\n",
      "hi\n",
      "(2, 19999)\n",
      "Computed 2D KDE\n",
      "min_density=0.0\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 10.0)\n",
      "hi\n",
      "(2, 19999)\n",
      "Computed 2D KDE\n",
      "min_density=0.0\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 10.0)\n",
      "hi\n",
      "(2, 19999)\n",
      "Computed 2D KDE\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-92-e789e39812dc>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      7\u001b[0m            \u001b[0maxes\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdata_j\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtitles_j\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mreshape\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mFalse\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmax_tIC\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      8\u001b[0m            \u001b[0mmin_density\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmax_diff\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m10.\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmax_i\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcustom_lims\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcustom_lims\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mn_levels\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m11\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 9\u001b[1;33m            n_pts=50j, parallel=False, remake=True)\n\u001b[0m",
      "\u001b[1;32m/vspdata/vspcompute/home/clarezhu/software/conformation/jointplot_d3.py\u001b[0m in \u001b[0;36mjointplots\u001b[1;34m(data, save_dir, titles, main, refcoords, refcoords_j, axes, reshape, data_j, titles_j, max_tIC, min_density, custom_lims, custom_lims_j, max_diff, tpt_paths, tpt_paths_j, n_levels, worker_pool, parallel, n_pts, all_apo_data, remake, max_i, min_i)\u001b[0m\n\u001b[0;32m     66\u001b[0m   \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     67\u001b[0m     \u001b[1;32mfor\u001b[0m \u001b[0mi_j_tuple\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mi_j_tuples\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 68\u001b[1;33m       \u001b[0mjointplot_partial\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mi_j_tuple\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     69\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     70\u001b[0m   \u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"Done plotting columns\"\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m/vspdata/vspcompute/home/clarezhu/software/conformation/jointplot_d3.py\u001b[0m in \u001b[0;36mjointplot\u001b[1;34m(i_j_tuple, all_data, save_dir, make_animation, trajectory, video_file, titles, main, include_1d_kde, custom_lims, custom_lims_j, axes, data_j, titles_j, refcoords, refcoords_j, max_tIC, min_density, max_diff, tpt_paths, tpt_paths_j, n_levels, n_pts, all_apo_data, remake, min_i, superpose_circles, cmap)\u001b[0m\n\u001b[0;32m    144\u001b[0m     \u001b[1;31m# Evaluate the KDE on a regular grid...\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    145\u001b[0m     \u001b[0mcoords\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mvstack\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mitem\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mravel\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mitem\u001b[0m \u001b[1;32min\u001b[0m \u001b[1;33m[\u001b[0m\u001b[0mxi\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0myi\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 146\u001b[1;33m     \u001b[0mdensity\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mkde\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mcoords\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mreshape\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mxi\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    147\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mall_apo_data\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    148\u001b[0m       \u001b[0mapo_density\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mapo_kde\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mcoords\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mreshape\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mxi\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m/home/enf/anaconda3/lib/python3.5/site-packages/scipy/stats/kde.py\u001b[0m in \u001b[0;36mevaluate\u001b[1;34m(self, points)\u001b[0m\n\u001b[0;32m    217\u001b[0m             \u001b[1;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mm\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    218\u001b[0m                 \u001b[0mdiff\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdataset\u001b[0m \u001b[1;33m-\u001b[0m \u001b[0mpoints\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mi\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mnewaxis\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 219\u001b[1;33m                 \u001b[0mtdiff\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mdot\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0minv_cov\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdiff\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    220\u001b[0m                 \u001b[0menergy\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0msum\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdiff\u001b[0m \u001b[1;33m*\u001b[0m \u001b[0mtdiff\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m/\u001b[0m \u001b[1;36m2.0\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    221\u001b[0m                 \u001b[0mresult\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mi\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0msum\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mexp\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m-\u001b[0m\u001b[0menergy\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m: "
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9d1c61aef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWsAAAEfCAYAAACH9eevAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJKyFECILSFgLIUQW+H8jkkxzr2uYiAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9d1f23e9b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWsAAAEfCAYAAACH9eevAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nB0hYCyFEDpCwFkKIHCBhLYQQOeD/A7anWosqf5xaAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9d1f23ee48>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import jointplot_d3\n",
    "reload(jointplot_d3)\n",
    "from jointplot_d3 import *\n",
    "jointplots(salm_tica_resampled[::1,:], \n",
    "           analysis_dir, titles = [\"tIC %d\" %i for i in range(1, n_components+1)], \n",
    "           main = \"mOR_salm_MSM_%d_Clusters\" %n_clusters, refcoords = ref_pp_tica, \n",
    "           axes=None, data_j=None, titles_j=None, reshape=False, max_tIC=1, \n",
    "           min_density=None, max_diff=10., max_i = 1, custom_lims=custom_lims, n_levels=11, \n",
    "           n_pts=50j, parallel=False, remake=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Making delta G plots.\n",
      "hi\n",
      "(2, 20003)\n",
      "Computed 2D KDE\n",
      "min_density=0.0\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 10.0)\n",
      "hi\n",
      "(2, 20003)\n",
      "Computed 2D KDE\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-93-ccb96876d80f>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      7\u001b[0m            \u001b[0maxes\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdata_j\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtitles_j\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mreshape\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mFalse\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmax_tIC\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      8\u001b[0m            \u001b[0mmin_density\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmax_diff\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m10.\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmax_i\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcustom_lims\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcustom_lims\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mn_levels\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m11\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 9\u001b[1;33m            n_pts=50j, parallel=False, remake=True)\n\u001b[0m",
      "\u001b[1;32m/vspdata/vspcompute/home/clarezhu/software/conformation/jointplot_d3.py\u001b[0m in \u001b[0;36mjointplots\u001b[1;34m(data, save_dir, titles, main, refcoords, refcoords_j, axes, reshape, data_j, titles_j, max_tIC, min_density, custom_lims, custom_lims_j, max_diff, tpt_paths, tpt_paths_j, n_levels, worker_pool, parallel, n_pts, all_apo_data, remake, max_i, min_i)\u001b[0m\n\u001b[0;32m     66\u001b[0m   \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     67\u001b[0m     \u001b[1;32mfor\u001b[0m \u001b[0mi_j_tuple\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mi_j_tuples\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 68\u001b[1;33m       \u001b[0mjointplot_partial\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mi_j_tuple\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     69\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     70\u001b[0m   \u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"Done plotting columns\"\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m/vspdata/vspcompute/home/clarezhu/software/conformation/jointplot_d3.py\u001b[0m in \u001b[0;36mjointplot\u001b[1;34m(i_j_tuple, all_data, save_dir, make_animation, trajectory, video_file, titles, main, include_1d_kde, custom_lims, custom_lims_j, axes, data_j, titles_j, refcoords, refcoords_j, max_tIC, min_density, max_diff, tpt_paths, tpt_paths_j, n_levels, n_pts, all_apo_data, remake, min_i, superpose_circles, cmap)\u001b[0m\n\u001b[0;32m    144\u001b[0m     \u001b[1;31m# Evaluate the KDE on a regular grid...\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    145\u001b[0m     \u001b[0mcoords\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mvstack\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mitem\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mravel\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mitem\u001b[0m \u001b[1;32min\u001b[0m \u001b[1;33m[\u001b[0m\u001b[0mxi\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0myi\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 146\u001b[1;33m     \u001b[0mdensity\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mkde\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mcoords\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mreshape\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mxi\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    147\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mall_apo_data\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    148\u001b[0m       \u001b[0mapo_density\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mapo_kde\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mcoords\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mreshape\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mxi\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m/home/enf/anaconda3/lib/python3.5/site-packages/scipy/stats/kde.py\u001b[0m in \u001b[0;36mevaluate\u001b[1;34m(self, points)\u001b[0m\n\u001b[0;32m    217\u001b[0m             \u001b[1;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mm\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    218\u001b[0m                 \u001b[0mdiff\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdataset\u001b[0m \u001b[1;33m-\u001b[0m \u001b[0mpoints\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mi\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mnewaxis\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 219\u001b[1;33m                 \u001b[0mtdiff\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mdot\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0minv_cov\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdiff\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    220\u001b[0m                 \u001b[0menergy\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0msum\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdiff\u001b[0m \u001b[1;33m*\u001b[0m \u001b[0mtdiff\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m/\u001b[0m \u001b[1;36m2.0\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    221\u001b[0m                 \u001b[0mresult\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mi\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0msum\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mexp\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m-\u001b[0m\u001b[0menergy\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m: "
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9d21723470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWsAAAEfCAYAAACH9eevAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9d21724e80>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import jointplot_d3\n",
    "reload(jointplot_d3)\n",
    "from jointplot_d3 import *\n",
    "jointplots(apo_tica_resampled[::1,:], \n",
    "           analysis_dir, titles = [\"tIC %d\" %i for i in range(1, n_components+1)], \n",
    "           main = \"apo_MSM_%d_Clusters\" %n_clusters, refcoords = ref_pp_tica, \n",
    "           axes=None, data_j=None, titles_j=None, reshape=False, max_tIC=1, \n",
    "           min_density=None, max_diff=10., max_i = 1, custom_lims=custom_lims, n_levels=11, \n",
    "           n_pts=50j, parallel=False, remake=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Making delta G plots.\n",
      "hi\n",
      "(2, 20000)\n",
      "Computed 2D KDE\n",
      "min_density=0.0\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 10.0)\n",
      "hi\n",
      "(2, 20000)\n",
      "Computed 2D KDE\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-94-0919f50ad803>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      7\u001b[0m            \u001b[0maxes\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdata_j\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtitles_j\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mreshape\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mFalse\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmax_tIC\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      8\u001b[0m            \u001b[0mmin_density\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmax_diff\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m10.\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmax_i\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcustom_lims\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcustom_lims\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mn_levels\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m11\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 9\u001b[1;33m            n_pts=50j, parallel=False, remake=True)\n\u001b[0m",
      "\u001b[1;32m/vspdata/vspcompute/home/clarezhu/software/conformation/jointplot_d3.py\u001b[0m in \u001b[0;36mjointplots\u001b[1;34m(data, save_dir, titles, main, refcoords, refcoords_j, axes, reshape, data_j, titles_j, max_tIC, min_density, custom_lims, custom_lims_j, max_diff, tpt_paths, tpt_paths_j, n_levels, worker_pool, parallel, n_pts, all_apo_data, remake, max_i, min_i)\u001b[0m\n\u001b[0;32m     66\u001b[0m   \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     67\u001b[0m     \u001b[1;32mfor\u001b[0m \u001b[0mi_j_tuple\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mi_j_tuples\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 68\u001b[1;33m       \u001b[0mjointplot_partial\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mi_j_tuple\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     69\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     70\u001b[0m   \u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"Done plotting columns\"\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m/vspdata/vspcompute/home/clarezhu/software/conformation/jointplot_d3.py\u001b[0m in \u001b[0;36mjointplot\u001b[1;34m(i_j_tuple, all_data, save_dir, make_animation, trajectory, video_file, titles, main, include_1d_kde, custom_lims, custom_lims_j, axes, data_j, titles_j, refcoords, refcoords_j, max_tIC, min_density, max_diff, tpt_paths, tpt_paths_j, n_levels, n_pts, all_apo_data, remake, min_i, superpose_circles, cmap)\u001b[0m\n\u001b[0;32m    144\u001b[0m     \u001b[1;31m# Evaluate the KDE on a regular grid...\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    145\u001b[0m     \u001b[0mcoords\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mvstack\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mitem\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mravel\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mitem\u001b[0m \u001b[1;32min\u001b[0m \u001b[1;33m[\u001b[0m\u001b[0mxi\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0myi\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 146\u001b[1;33m     \u001b[0mdensity\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mkde\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mcoords\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mreshape\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mxi\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    147\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mall_apo_data\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    148\u001b[0m       \u001b[0mapo_density\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mapo_kde\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mcoords\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mreshape\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mxi\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m/home/enf/anaconda3/lib/python3.5/site-packages/scipy/stats/kde.py\u001b[0m in \u001b[0;36mevaluate\u001b[1;34m(self, points)\u001b[0m\n\u001b[0;32m    217\u001b[0m             \u001b[1;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mm\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    218\u001b[0m                 \u001b[0mdiff\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdataset\u001b[0m \u001b[1;33m-\u001b[0m \u001b[0mpoints\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mi\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mnewaxis\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 219\u001b[1;33m                 \u001b[0mtdiff\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mdot\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0minv_cov\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdiff\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    220\u001b[0m                 \u001b[0menergy\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0msum\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdiff\u001b[0m \u001b[1;33m*\u001b[0m \u001b[0mtdiff\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m/\u001b[0m \u001b[1;36m2.0\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    221\u001b[0m                 \u001b[0mresult\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mi\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0msum\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mexp\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m-\u001b[0m\u001b[0menergy\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m: "
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9d21869898>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWsAAAEfCAYAAACH9eevAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f9d2057d438>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import jointplot_d3\n",
    "reload(jointplot_d3)\n",
    "from jointplot_d3 import *\n",
    "jointplots(salm_nb_tica_resampled[::1,:], \n",
    "           analysis_dir, titles = [\"tIC %d\" %i for i in range(1, n_components+1)], \n",
    "           main = \"salm_nb_MSM_%d_Clusters\" %n_clusters, refcoords = ref_pp_tica, \n",
    "           axes=None, data_j=None, titles_j=None, reshape=False, max_tIC=1, \n",
    "           min_density=None, max_diff=10., max_i = 1, custom_lims=custom_lims, n_levels=11, \n",
    "           n_pts=50j, parallel=False, remake=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "jointplots(salm_nb_pnas_resampled[::1,:], analysis_dir, titles = sorted(feature_name_residues_dict.keys()), main = \"mOR_salm_nb_MSM_%d_Clusters\" %n_clusters, refcoords=ref_pnas, axes=None, data_j=None, titles_j=None, reshape=False, max_tIC=20, min_density=None, max_diff=3., custom_lims=custom_lims, n_levels=10, remake=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "tica_resampled = np.vstack([compat_verboseload(salm_tica_resampled_file)[::1,], compat_verboseload(apo_tica_resampled_file)[::1,], compat_verboseload(salm_nb_tica_resampled_file)[::1,]])\n",
    "mins = np.min(tica_resampled, axis=0)\n",
    "maxs = np.max(tica_resampled, axis=0)\n",
    "stds = np.std(tica_resampled, axis=0)\n",
    "custom_lims = [[mins[i] - 1.*stds[i], maxs[i] + 1.*stds[i]] for i in range(0,len(mins))]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "custom_lims"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "print(custom_lims)\n",
    "print(custom_lims[3][1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Making delta G plots.\n",
      "1\n",
      "2\n",
      "3\n",
      "4\n",
      "5\n",
      "6\n",
      "(2, 19984)\n",
      "Computed 2D KDE\n",
      "min_density=0.0\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732   0.73737192]\n",
      " [-0.27086734  0.45005798]]\n",
      "7\n",
      "(2, 19984)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732  -2.846076  ]\n",
      " [-0.27086734  0.34059189]]\n",
      "8\n",
      "(2, 19984)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732  -0.77531983]\n",
      " [-0.27086734  0.15601567]]\n",
      "9\n",
      "(2, 19984)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732  -0.86493743]\n",
      " [-0.27086734  0.67270475]]\n",
      "Done plotting columns\n",
      "Making delta G plots.\n",
      "1\n",
      "(2, 19998)\n",
      "Computed 2D KDE\n",
      "min_density=0.0\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732   1.16827855]\n",
      " [-0.27086734 -0.90431764]]\n",
      "2\n",
      "(2, 19998)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732  -0.12336551]\n",
      " [-0.27086734 -0.51214134]]\n",
      "3\n",
      "(2, 19998)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732  -1.66948062]\n",
      " [-0.27086734  0.09047833]]\n",
      "4\n",
      "(2, 19998)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732   1.57437074]\n",
      " [-0.27086734 -0.16239742]]\n",
      "5\n",
      "(2, 19998)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732   0.43695591]\n",
      " [-0.27086734 -0.17755717]]\n",
      "6\n",
      "(2, 19998)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732   0.73737192]\n",
      " [-0.27086734  0.45005798]]\n",
      "7\n",
      "(2, 19998)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732  -2.846076  ]\n",
      " [-0.27086734  0.34059189]]\n",
      "8\n",
      "(2, 19998)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732  -0.77531983]\n",
      " [-0.27086734  0.15601567]]\n",
      "9\n",
      "(2, 19998)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732  -0.86493743]\n",
      " [-0.27086734  0.67270475]]\n",
      "Done plotting columns\n",
      "Making delta G plots.\n",
      "1\n",
      "(2, 19994)\n",
      "Computed 2D KDE\n",
      "min_density=0.0\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732   1.16827855]\n",
      " [-0.27086734 -0.90431764]]\n",
      "2\n",
      "(2, 19994)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732  -0.12336551]\n",
      " [-0.27086734 -0.51214134]]\n",
      "3\n",
      "(2, 19994)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732  -1.66948062]\n",
      " [-0.27086734  0.09047833]]\n",
      "4\n",
      "(2, 19994)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732   1.57437074]\n",
      " [-0.27086734 -0.16239742]]\n",
      "5\n",
      "(2, 19994)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732   0.43695591]\n",
      " [-0.27086734 -0.17755717]]\n",
      "6\n",
      "(2, 19994)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732   0.73737192]\n",
      " [-0.27086734  0.45005798]]\n",
      "7\n",
      "(2, 19994)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732  -2.846076  ]\n",
      " [-0.27086734  0.34059189]]\n",
      "8\n",
      "(2, 19994)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732  -0.77531983]\n",
      " [-0.27086734  0.15601567]]\n",
      "9\n",
      "(2, 19994)\n",
      "Computed 2D KDE\n",
      "Computed Density. Now plotting.\n",
      "(0.0, 8.0)\n",
      "[[ 0.1997732  -0.86493743]\n",
      " [-0.27086734  0.67270475]]\n",
      "Done plotting columns\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd536655160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd5340a04e0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd53657e5c0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd5340a0da0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd52fee2eb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd53655d9e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd52c1c2198>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd52c1273c8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd5228a64a8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd5229cae80>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd5314d5550>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd5314d5dd8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd52c220390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd52c220a20>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd536638668>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd53655db70>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd53416da20>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd52c248748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd5228c30f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd533ed54e0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd522904a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd5228fa1d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd52c0a29b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import jointplot_d3\n",
    "reload(jointplot_d3)\n",
    "from jointplot_d3 import *\n",
    "refcoords_jointplots_dir = \"%s/refcoords_jointplots\" %analysis_dir\n",
    "if not os.path.exists(refcoords_jointplots_dir):\n",
    "    os.makedirs(refcoords_jointplots_dir)\n",
    "jointplots(compat_verboseload(salm_tica_resampled_file)[::1,:], refcoords_jointplots_dir, titles = [\"tIC%d\" %j for j in range(1,n_components+1)], main = \"salm MSM\", refcoords=ref_pp_tica, axes=None, data_j=None, titles_j=None, reshape=False, max_tIC=10, min_density=None, max_diff=8., custom_lims=custom_lims, n_pts=200j, n_levels=15, remake=False, parallel=False, max_i=1)\n",
    "jointplots(compat_verboseload(apo_tica_resampled_file)[::1,:], refcoords_jointplots_dir, titles = [\"tIC%d\" %j for j in range(1,n_components+1)], main = \"Apo MSM\", refcoords=ref_pp_tica, axes=None, data_j=None, titles_j=None, reshape=False, max_tIC=10, min_density=None, max_diff=8., custom_lims=custom_lims, n_pts=200j, n_levels=15, remake=False, parallel=False, max_i=1)\n",
    "jointplots(compat_verboseload(salm_nb_tica_resampled_file)[::1,:], refcoords_jointplots_dir, titles = [\"tIC%d\" %j for j in range(1,n_components+1)], main = \"salm_nbentanil MSM\", refcoords=ref_pp_tica, axes=None, data_j=None, titles_j=None, reshape=False, max_tIC=10, min_density=None, max_diff=8., custom_lims=custom_lims, n_pts=200j, n_levels=15, remake=False, parallel=False, max_i=1)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "ref_pp_tica"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "np.concatenate(salm_tica).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "np.concatenate(apo_tica).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import jointplot_d3\n",
    "reload(jointplot_d3)\n",
    "from jointplot_d3 import *\n",
    "jointplots(compat_verboseload(salm_tica_resampled_file)[::1,:], analysis_dir, titles = [\"tIC%d\" %j for j in range(1,n_components+1)], main = \"salm MSM\", refcoords=None, axes=None, data_j=salm_pnas_resampled, titles_j=sorted(feature_name_residues_dict.keys()), reshape=False, max_tIC=10, min_density=None, max_diff=8., custom_lims=None, remake=True)\n",
    "#jointplots(compat_verboseload(apo_tica_resampled_file)[::1,:], analysis_dir, titles = [\"tIC%d\" %j for j in range(1,n_components+1)], main = \"Apo MSM\", refcoords=None, axes=None, data_j=apo_pnas_resampled, titles_j=sorted(feature_name_residues_dict.keys()), reshape=False, max_tIC=10, min_density=None, max_diff=8., custom_lims=None)\n",
    "#jointplots(compat_verboseload(salm_nb_tica_resampled_file)[::1,:], analysis_dir, titles = [\"tIC%d\" %j for j in range(1,n_components+1)], main = \"salm_nbentanil MSM\", refcoords=None, axes=None, data_j=salm_pnas_resampled, titles_j=sorted(feature_name_residues_dict.keys()), reshape=False, max_tIC=10, min_density=None, max_diff=8., custom_lims=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "import detect_intermediates\n",
    "reload(detect_intermediates)\n",
    "from detect_intermediates import *\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#sample_tIC_regions_silverman(np.concatenate(compat_verboseload(projected_features_dir))[::100,:], analysis_dir)\n",
    "#get_kde_mins_and_maxes(np.concatenate(compat_verboseload(projected_features_dir))[::100,:], analysis_dir)\n",
    "sample_kde_maxima(compat_verboseload(projected_features_dir), analysis_dir, trajs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "jointplots(compat_verboseload(apo_tica_resampled_file)[::1,:], analysis_dir, titles = [\"tIC%d\" %j for j in range(1,n_components+1)], main = \"Apo MSM\", refcoords_file = None, axes=None, data_j=None, titles_j=None, reshape=False, max_tIC=5, min_density=None, max_diff=5., custom_lims=custom_lims)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "plot_columns(tica_dir, salm_pnas_resampled_file, titles = sorted(feature_name_residues_dict.keys()), main = \"mOR_salm_MSM_\", tICA = False, scale = 1.0, refcoords_file = None, axes=None, concatenate=False, reshape=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#salm_tica = compat_verboseload(salm_tica_file)\n",
    "#salm_tica_resampled_file = os.path.join(tica_dir, \"salm_tica_msm_lag-time%d_prior-counts%s_clusters%d_resampled.h5\" %(lag_time, str(prior_counts), n_clusters))\n",
    "#resample_features_by_msm_equilibirum_pop(salm_tica, salm_resampled_traj_to_frames, salm_tica_resampled_file)\n",
    "#salm_tica_resampled = compat_verboseload(salm_tica_resampled_file)\n",
    "plot_columns(tica_dir, salm_tica_resampled_file, titles = [\"tIC.%d\" %i for i in range(1,n_components+1)], main = \"mOR_salm_MSM_\", tICA = False, scale = 1.0, refcoords_file = ref_tica_coords, axes=tica_axes, concatenate=False, reshape=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "salm_clusterer = compat_verboseload(salm_clusterer_file)\n",
    "salm_msm = compat_verboseload(salm_msm_dir)\n",
    "\n",
    "cluster_id_proportion_tuples = []\n",
    "\n",
    "equilibrium_populations = salm_msm.populations_\n",
    "\n",
    "for cluster_id in salm_msm.mapping_.keys():\n",
    "    state_id = salm_msm.mapping_[cluster_id]\n",
    "    cluster_id_proportion_tuples.append((cluster_id,equilibrium_populations[state_id]))\n",
    "cluster_id_proportion_tuples = sorted(cluster_id_proportion_tuples, key=itemgetter(1), reverse=True)\n",
    "\n",
    "#plot_all_tics_and_clusters(tica_dir, salm_tica_resampled_file, salm_clusterer_file, lag_time, tic_range=[0,3], main = \"salm_MSM_Reweighted\", label = \"cluster_id\", active_cluster_ids = [c[0] for c in cluster_id_proportion_tuples[0:100]], intermediate_cluster_ids = [], inactive_cluster_ids = [], concatenate=False, axes=tica_axes)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "len(dview)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/bu72_clusterer_1000clusters_1samples_samples_kdtree\n",
      "loading \"/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/bu72_clusterer1000.h5\"...\n",
      "loading \"/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/bu72_tica.h5\"...\n",
      "loading \"/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/bu72_clusterer1000.h5\"...\n",
      "[[149 356]\n",
      " [379 392]\n",
      " [468 474]\n",
      " ..., \n",
      " [218 185]\n",
      " [375 365]\n",
      " [294 237]]\n",
      "1000\n",
      "Saving \"/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/bu72_clusterer_1000clusters_1samples_samples_kdtree_indices.h5\"... (<class 'numpy.ndarray'>)\n"
     ]
    }
   ],
   "source": [
    "import custom_clusterer\n",
    "reload(custom_clusterer)\n",
    "from custom_clusterer import *\n",
    "\n",
    "n_samples = 1\n",
    "n_clusters= 1000\n",
    "samples_dir = \"%s/salm_clusterer_%dclusters_%dsamples_samples_kdtree\" %(tica_dir, n_clusters, n_samples)\n",
    "samples_indices_file = \"%s/salm_clusterer_%dclusters_%dsamples_samples_kdtree_indices.h5\" %(tica_dir, n_clusters, n_samples)\n",
    "if not os.path.exists(samples_dir): os.makedirs(samples_dir)\n",
    "print(samples_dir)\n",
    "traj_files = get_trajectory_files(traj_dir, traj_ext)    \n",
    "sample_from_clusterer(salm_clusterer_file, salm_tica_file, [traj_files[i] for i in salm_trajs], \n",
    "                      n_samples, samples_dir, samples_indices_file,\n",
    "                      worker_pool=dview, parallel=True, lig_name=\"LIG\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/suf_clusterer_1000clusters_1samples_samples_kdtree\n",
      "loading \"/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/suf_clusterer1000.h5\"...\n",
      "loading \"/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/suf_tica.h5\"...\n",
      "loading \"/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/suf_clusterer1000.h5\"...\n",
      "[[494 114]\n",
      " [218 323]\n",
      " [ 38 170]\n",
      " ..., \n",
      " [726 453]\n",
      " [714  48]\n",
      " [615 459]]\n",
      "1000\n",
      "Saving \"/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/suf_clusterer_1000clusters_1samples_samples_kdtree_indices.h5\"... (<class 'numpy.ndarray'>)\n"
     ]
    }
   ],
   "source": [
    "import custom_clusterer\n",
    "reload(custom_clusterer)\n",
    "from custom_clusterer import *\n",
    "\n",
    "n_samples = 1\n",
    "n_clusters= 1000\n",
    "samples_dir = \"%s/salm_nb_clusterer_%dclusters_%dsamples_samples_kdtree\" %(tica_dir, n_clusters, n_samples)\n",
    "samples_indices_file = \"%s/salm_nb_clusterer_%dclusters_%dsamples_samples_kdtree_indices.h5\" %(tica_dir, n_clusters, n_samples)\n",
    "if not os.path.exists(samples_dir): os.makedirs(samples_dir)\n",
    "print(samples_dir)\n",
    "traj_files = get_trajectory_files(traj_dir, traj_ext)    \n",
    "sample_from_clusterer(salm_nb_clusterer_file, salm_nb_tica_file, [traj_files[i] for i in salm_nb_trajs], \n",
    "                      n_samples, samples_dir, samples_indices_file,\n",
    "                      worker_pool=dview, parallel=True, lig_name=\"LIG\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/all_clusterer_100clusters_1samples_samples_kdtree\n",
      "loading \"/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/clusterer_100clusters.h5\"...\n",
      "loading \"/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/phi_psi_chi2_allprot_projected.h5\"...\n",
      "loading \"/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/clusterer_100clusters.h5\"...\n",
      "[[ 965    5]\n",
      " [ 760  289]\n",
      " [1444   39]\n",
      " [1667  224]\n",
      " [ 485   54]\n",
      " [ 339  428]\n",
      " [ 669  179]\n",
      " [ 783  303]\n",
      " [  99   47]\n",
      " [ 835   30]\n",
      " [1027  332]\n",
      " [ 354  101]\n",
      " [1708    7]\n",
      " [1663  356]\n",
      " [1591   58]\n",
      " [ 222  135]\n",
      " [1560  160]\n",
      " [1357   23]\n",
      " [ 534  242]\n",
      " [1181  374]\n",
      " [ 323  535]\n",
      " [ 854  521]\n",
      " [ 419   70]\n",
      " [1770  172]\n",
      " [ 303  470]\n",
      " [1516   76]\n",
      " [ 896  148]\n",
      " [ 885  521]\n",
      " [ 335  218]\n",
      " [1178  225]\n",
      " [1543   64]\n",
      " [1669  513]\n",
      " [ 586  274]\n",
      " [ 379  461]\n",
      " [1628  372]\n",
      " [ 992  211]\n",
      " [ 271  551]\n",
      " [ 161   11]\n",
      " [1602  499]\n",
      " [1609  188]\n",
      " [ 343   74]\n",
      " [ 309  436]\n",
      " [ 284  366]\n",
      " [ 813  437]\n",
      " [  64  375]\n",
      " [1707  137]\n",
      " [1592  482]\n",
      " [ 338   95]\n",
      " [1051  262]\n",
      " [1726  371]\n",
      " [1112  266]\n",
      " [  92  148]\n",
      " [1528  241]\n",
      " [1741  274]\n",
      " [ 781  532]\n",
      " [1241   24]\n",
      " [1730  572]\n",
      " [ 127  293]\n",
      " [ 210  165]\n",
      " [1476   10]\n",
      " [ 971   19]\n",
      " [ 805  429]\n",
      " [1761  383]\n",
      " [ 274  543]\n",
      " [ 210  220]\n",
      " [ 809  200]\n",
      " [  66  367]\n",
      " [ 400  221]\n",
      " [ 316  102]\n",
      " [ 255  304]\n",
      " [1241  332]\n",
      " [1631    0]\n",
      " [ 714   32]\n",
      " [  10  231]\n",
      " [1647  270]\n",
      " [1556  263]\n",
      " [  26  460]\n",
      " [ 873  328]\n",
      " [1006  231]\n",
      " [1089    3]\n",
      " [ 321  171]\n",
      " [1549  270]\n",
      " [ 372  220]\n",
      " [1677   19]\n",
      " [ 980  426]\n",
      " [1078  179]\n",
      " [1084  404]\n",
      " [1184   30]\n",
      " [1519   15]\n",
      " [1747  249]\n",
      " [1368   66]\n",
      " [1610  223]\n",
      " [1015   58]\n",
      " [ 453   11]\n",
      " [1691  408]\n",
      " [1709  213]\n",
      " [ 292   14]\n",
      " [1637  473]\n",
      " [1708  526]\n",
      " [1248    8]]\n",
      "100\n",
      "()\n",
      "Converting to list\n",
      "[965, 5, 0, 0]\n",
      "965\n",
      "()\n",
      "Converting to list\n",
      "[760, 289, 1, 0]\n",
      "760\n",
      "()\n",
      "Converting to list\n",
      "[1444, 39, 2, 0]\n",
      "1444\n",
      "()\n",
      "Converting to list\n",
      "[1667, 224, 3, 0]\n",
      "1667\n",
      "()\n",
      "Converting to list\n",
      "[485, 54, 4, 0]\n",
      "485\n",
      "()\n",
      "Converting to list\n",
      "[339, 428, 5, 0]\n",
      "339\n",
      "()\n",
      "Converting to list\n",
      "[669, 179, 6, 0]\n",
      "669\n",
      "()\n",
      "Converting to list\n",
      "[783, 303, 7, 0]\n",
      "783\n",
      "()\n",
      "Converting to list\n",
      "[99, 47, 8, 0]\n",
      "99\n",
      "()\n",
      "Converting to list\n",
      "[835, 30, 9, 0]\n",
      "835\n",
      "()\n",
      "Converting to list\n",
      "[1027, 332, 10, 0]\n",
      "1027\n",
      "()\n",
      "Converting to list\n",
      "[354, 101, 11, 0]\n",
      "354\n",
      "()\n",
      "Converting to list\n",
      "[1708, 7, 12, 0]\n",
      "1708\n",
      "()\n",
      "Converting to list\n",
      "[1663, 356, 13, 0]\n",
      "1663\n",
      "()\n",
      "Converting to list\n",
      "[1591, 58, 14, 0]\n",
      "1591\n",
      "()\n",
      "Converting to list\n",
      "[222, 135, 15, 0]\n",
      "222\n",
      "()\n",
      "Converting to list\n",
      "[1560, 160, 16, 0]\n",
      "1560\n",
      "()\n",
      "Converting to list\n",
      "[1357, 23, 17, 0]\n",
      "1357\n",
      "()\n",
      "Converting to list\n",
      "[534, 242, 18, 0]\n",
      "534\n",
      "()\n",
      "Converting to list\n",
      "[1181, 374, 19, 0]\n",
      "1181\n",
      "()\n",
      "Converting to list\n",
      "[323, 535, 20, 0]\n",
      "323\n",
      "()\n",
      "Converting to list\n",
      "[854, 521, 21, 0]\n",
      "854\n",
      "()\n",
      "Converting to list\n",
      "[419, 70, 22, 0]\n",
      "419\n",
      "()\n",
      "Converting to list\n",
      "[1770, 172, 23, 0]\n",
      "1770\n",
      "()\n",
      "Converting to list\n",
      "[303, 470, 24, 0]\n",
      "303\n",
      "()\n",
      "Converting to list\n",
      "[1516, 76, 25, 0]\n",
      "1516\n",
      "()\n",
      "Converting to list\n",
      "[896, 148, 26, 0]\n",
      "896\n",
      "()\n",
      "Converting to list\n",
      "[885, 521, 27, 0]\n",
      "885\n",
      "()\n",
      "Converting to list\n",
      "[335, 218, 28, 0]\n",
      "335\n",
      "()\n",
      "Converting to list\n",
      "[1178, 225, 29, 0]\n",
      "1178\n",
      "()\n",
      "Converting to list\n",
      "[1543, 64, 30, 0]\n",
      "1543\n",
      "()\n",
      "Converting to list\n",
      "[1669, 513, 31, 0]\n",
      "1669\n",
      "()\n",
      "Converting to list\n",
      "[586, 274, 32, 0]\n",
      "586\n",
      "()\n",
      "Converting to list\n",
      "[379, 461, 33, 0]\n",
      "379\n",
      "()\n",
      "Converting to list\n",
      "[1628, 372, 34, 0]\n",
      "1628\n",
      "()\n",
      "Converting to list\n",
      "[992, 211, 35, 0]\n",
      "992\n",
      "()\n",
      "Converting to list\n",
      "[271, 551, 36, 0]\n",
      "271\n",
      "()\n",
      "Converting to list\n",
      "[161, 11, 37, 0]\n",
      "161\n",
      "()\n",
      "Converting to list\n",
      "[1602, 499, 38, 0]\n",
      "1602\n",
      "()\n",
      "Converting to list\n",
      "[1609, 188, 39, 0]\n",
      "1609\n",
      "()\n",
      "Converting to list\n",
      "[343, 74, 40, 0]\n",
      "343\n",
      "()\n",
      "Converting to list\n",
      "[309, 436, 41, 0]\n",
      "309\n",
      "()\n",
      "Converting to list\n",
      "[284, 366, 42, 0]\n",
      "284\n",
      "()\n",
      "Converting to list\n",
      "[813, 437, 43, 0]\n",
      "813\n",
      "()\n",
      "Converting to list\n",
      "[64, 375, 44, 0]\n",
      "64\n",
      "()\n",
      "Converting to list\n",
      "[1707, 137, 45, 0]\n",
      "1707\n",
      "()\n",
      "Converting to list\n",
      "[1592, 482, 46, 0]\n",
      "1592\n",
      "()\n",
      "Converting to list\n",
      "[338, 95, 47, 0]\n",
      "338\n",
      "()\n",
      "Converting to list\n",
      "[1051, 262, 48, 0]\n",
      "1051\n",
      "()\n",
      "Converting to list\n",
      "[1726, 371, 49, 0]\n",
      "1726\n",
      "()\n",
      "Converting to list\n",
      "[1112, 266, 50, 0]\n",
      "1112\n",
      "()\n",
      "Converting to list\n",
      "[92, 148, 51, 0]\n",
      "92\n",
      "()\n",
      "Converting to list\n",
      "[1528, 241, 52, 0]\n",
      "1528\n",
      "()\n",
      "Converting to list\n",
      "[1741, 274, 53, 0]\n",
      "1741\n",
      "()\n",
      "Converting to list\n",
      "[781, 532, 54, 0]\n",
      "781\n",
      "()\n",
      "Converting to list\n",
      "[1241, 24, 55, 0]\n",
      "1241\n",
      "()\n",
      "Converting to list\n",
      "[1730, 572, 56, 0]\n",
      "1730\n",
      "()\n",
      "Converting to list\n",
      "[127, 293, 57, 0]\n",
      "127\n",
      "()\n",
      "Converting to list\n",
      "[210, 165, 58, 0]\n",
      "210\n",
      "()\n",
      "Converting to list\n",
      "[1476, 10, 59, 0]\n",
      "1476\n",
      "()\n",
      "Converting to list\n",
      "[971, 19, 60, 0]\n",
      "971\n",
      "()\n",
      "Converting to list\n",
      "[805, 429, 61, 0]\n",
      "805\n",
      "()\n",
      "Converting to list\n",
      "[1761, 383, 62, 0]\n",
      "1761\n",
      "()\n",
      "Converting to list\n",
      "[274, 543, 63, 0]\n",
      "274\n",
      "()\n",
      "Converting to list\n",
      "[210, 220, 64, 0]\n",
      "210\n",
      "()\n",
      "Converting to list\n",
      "[809, 200, 65, 0]\n",
      "809\n",
      "()\n",
      "Converting to list\n",
      "[66, 367, 66, 0]\n",
      "66\n",
      "()\n",
      "Converting to list\n",
      "[400, 221, 67, 0]\n",
      "400\n",
      "()\n",
      "Converting to list\n",
      "[316, 102, 68, 0]\n",
      "316\n",
      "()\n",
      "Converting to list\n",
      "[255, 304, 69, 0]\n",
      "255\n",
      "()\n",
      "Converting to list\n",
      "[1241, 332, 70, 0]\n",
      "1241\n",
      "()\n",
      "Converting to list\n",
      "[1631, 0, 71, 0]\n",
      "1631\n",
      "()\n",
      "Converting to list\n",
      "[714, 32, 72, 0]\n",
      "714\n",
      "()\n",
      "Converting to list\n",
      "[10, 231, 73, 0]\n",
      "10\n",
      "()\n",
      "Converting to list\n",
      "[1647, 270, 74, 0]\n",
      "1647\n",
      "()\n",
      "Converting to list\n",
      "[1556, 263, 75, 0]\n",
      "1556\n",
      "()\n",
      "Converting to list\n",
      "[26, 460, 76, 0]\n",
      "26\n",
      "()\n",
      "Converting to list\n",
      "[873, 328, 77, 0]\n",
      "873\n",
      "()\n",
      "Converting to list\n",
      "[1006, 231, 78, 0]\n",
      "1006\n",
      "()\n",
      "Converting to list\n",
      "[1089, 3, 79, 0]\n",
      "1089\n",
      "()\n",
      "Converting to list\n",
      "[321, 171, 80, 0]\n",
      "321\n",
      "()\n",
      "Converting to list\n",
      "[1549, 270, 81, 0]\n",
      "1549\n",
      "()\n",
      "Converting to list\n",
      "[372, 220, 82, 0]\n",
      "372\n",
      "()\n",
      "Converting to list\n",
      "[1677, 19, 83, 0]\n",
      "1677\n",
      "()\n",
      "Converting to list\n",
      "[980, 426, 84, 0]\n",
      "980\n",
      "()\n",
      "Converting to list\n",
      "[1078, 179, 85, 0]\n",
      "1078\n",
      "()\n",
      "Converting to list\n",
      "[1084, 404, 86, 0]\n",
      "1084\n",
      "()\n",
      "Converting to list\n",
      "[1184, 30, 87, 0]\n",
      "1184\n",
      "()\n",
      "Converting to list\n",
      "[1519, 15, 88, 0]\n",
      "1519\n",
      "()\n",
      "Converting to list\n",
      "[1747, 249, 89, 0]\n",
      "1747\n",
      "()\n",
      "Converting to list\n",
      "[1368, 66, 90, 0]\n",
      "1368\n",
      "()\n",
      "Converting to list\n",
      "[1610, 223, 91, 0]\n",
      "1610\n",
      "()\n",
      "Converting to list\n",
      "[1015, 58, 92, 0]\n",
      "1015\n",
      "()\n",
      "Converting to list\n",
      "[453, 11, 93, 0]\n",
      "453\n",
      "()\n",
      "Converting to list\n",
      "[1691, 408, 94, 0]\n",
      "1691\n",
      "()\n",
      "Converting to list\n",
      "[1709, 213, 95, 0]\n",
      "1709\n",
      "()\n",
      "Converting to list\n",
      "[292, 14, 96, 0]\n",
      "292\n",
      "()\n",
      "Converting to list\n",
      "[1637, 473, 97, 0]\n",
      "1637\n",
      "()\n",
      "Converting to list\n",
      "[1708, 526, 98, 0]\n",
      "1708\n",
      "()\n",
      "Converting to list\n",
      "[1248, 8, 99, 0]\n",
      "1248\n",
      "Saving \"/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/all_clusterer_100clusters_1samples_samples_kdtree_indices.h5\"... (<class 'numpy.ndarray'>)\n"
     ]
    }
   ],
   "source": [
    "import custom_clusterer\n",
    "reload(custom_clusterer)\n",
    "from custom_clusterer import *\n",
    "\n",
    "n_samples = 1\n",
    "n_clusters= 100\n",
    "samples_dir = \"%s/all_clusterer_%dclusters_%dsamples_samples_kdtree\" %(tica_dir, n_clusters, n_samples)\n",
    "samples_indices_file = \"%s/all_clusterer_%dclusters_%dsamples_samples_kdtree_indices.h5\" %(tica_dir, n_clusters, n_samples)\n",
    "if not os.path.exists(samples_dir): os.makedirs(samples_dir)\n",
    "print(samples_dir)\n",
    "traj_files = get_trajectory_files(traj_dir, traj_ext)    \n",
    "sample_from_clusterer(clusterer_dir, projected_features_dir, traj_files, \n",
    "                      n_samples, samples_dir, samples_indices_file,\n",
    "                      worker_pool=None, parallel=False, lig_name=\"NON\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/home/enf/md_simulations/MOR/h8_reimaged/sparse-tICA_t50_n_components10all_residues_4dkl_5c1m_under_cutoff6A-CA-py3-far_regularization_wolf_autoShrinkage0pt01-backup/all_clusterer_100clusters_1samples_samples_kdtree'"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "samples_dir"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "get_trajectory_files(traj_dir, traj_ext)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import custom_clusterer\n",
    "reload(custom_clusterer)\n",
    "from custom_clusterer import *\n",
    "\n",
    "n_samples = 5\n",
    "n_clusters=250\n",
    "samples_dir = \"%s/apo_clusterer_%dclusters_%dsamples_samples_kdtree\" %(tica_dir, n_clusters, n_samples)\n",
    "samples_indices_file = \"%s/apo_clusterer_%dclusters_%dsamples_samples_kdtree_indices.h5\" %(tica_dir, n_clusters, n_samples)\n",
    "if not os.path.exists(samples_dir): os.makedirs(samples_dir)\n",
    "\n",
    "sample_from_clusterer(apo_clusterer_file, apo_tica_file, [traj_files[i] for i in apo_trajs],\n",
    "                      n_samples, samples_dir, samples_indices_file,\n",
    "                      worker_pool=None, parallel=True, lig_name=\"LIG\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "traj_files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "feature_name_residues_dict.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "\n",
    "np.shape(apo_pnas_resampled)\n",
    "msm_dir = os.path.join(tica_dir, \"msm_lag-time%d_prior-counts%s_clusters%d_analysis\" %(lag_time, str(prior_counts), n_clusters))\n",
    "if not os.path.exists(msm_dir): os.makedirs(msm_dir)\n",
    "\n",
    "#plot_columns(msm_dir, apo_pnas_resampled_file, titles = feature_name_residues_dict.keys(), main = \"Apo-mOR\", tICA = False, scale = 1.0, refcoords_file = None, axes=None, concatenate=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import pip\n",
    "\n",
    "def install(package):\n",
    "    pip.main(['install', package])\n",
    "    \n",
    "install(\"moviepy\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import jointplot_d3\n",
    "reload(jointplot_d3)\n",
    "from jointplot_d3 import *\n",
    "\n",
    "for i, traj in enumerate(get_trajectory_files(traj_dir, traj_ext)):\n",
    "    if \"apo_rep_68.\" in traj:\n",
    "        index=i\n",
    "        break\n",
    "\n",
    "jointplot(apo_pnas_resampled[::10,[6,3]], \"%s/test.pdf\" %msm_dir, trajectory=apo_pnas[index][::1,[6,3]], \n",
    "          titles=[\"TM6 to TM3 distance\", \"RMSD of NPxxY to Inactive\"], main = \"mOR Activation Pathway\", \n",
    "          video_file = \"%s/apo_68.mp4\" %msm_dir, custom_xlim=[4.0, 16.0], custom_ylim= [0.0, 1.0])\n",
    "#test(\"%s/test2.mp4\" %msm_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "build_msm(salm_clusterer_file, lag_time, msm_model_dir)\n",
    "build_msm(apo_clusterer_file, lag_time, msm_model_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "tics_1_5_file = os.path.join(tica_dir, \"tICs_1_5.h5\")\n",
    "tics = load_file(projected_features_dir)\n",
    "tics_1_5 = [x[:,range(1,5)] for x in tics]\n",
    "verbosedump(tics_1_5, tics_1_5_file)\n",
    "\n",
    "clusterer300_clusters = os.path.join(tica_dir, \"clusterer300_clusters_tICs1_5.h5\")\n",
    "cluster_minikmeans(tica_dir, tics_1_5_file, traj_dir, n_clusters=300, clusterer_dir=clusterer300_clusters)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "len(dview)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import custom_clusterer\n",
    "reload(custom_clusterer)\n",
    "from custom_clusterer import *\n",
    "\n",
    "n_samples = 1\n",
    "n_clusters = 300\n",
    "\n",
    "traj_files = get_trajectory_files(traj_dir, traj_ext)\n",
    "featurized_trajectories = [traj_files[i] for i in feature_indices]\n",
    "salm_nb_traj_files = [t for t in featurized_trajectories if \"salm_nb\" in t]\n",
    "reduced_clusterer = compat_verboseload(clusterer300_clusters)\n",
    "salm_nb_reduced_clusterer_file = \"%s/salm_nb_reduced_clusterer.h5\" %tica_dir\n",
    "\n",
    "salm_nb_reduced_tica_file = \"%s/salm_nb_reduced_tica.h5\" %tica_dir\n",
    "salm_nb_reduced_tica = [x[:,range(1,5)] for x in salm_nb_tica]\n",
    "salm_nb_reseeded_dir = \"%s/salm_nb_reseeded_samples\" %tica_dir\n",
    "if not os.path.exists(salm_nb_reseeded_dir):\n",
    "    os.makedirs(salm_nb_reseeded_dir)\n",
    "verbosedump(salm_nb_reduced_tica, salm_nb_reduced_tica_file)\n",
    "\n",
    "salm_nb_reduced_clusters = [reduced_clusterer.labels_[i] for i in salm_nb_trajs]\n",
    "salm_nb_reduced_clusterer = copy.deepcopy(reduced_clusterer)\n",
    "salm_nb_reduced_clusterer.labels_ = salm_nb_reduced_clusters\n",
    "verbosedump(salm_nb_reduced_clusterer, salm_nb_reduced_clusterer_file)\n",
    "#reseed_from_clusterer(salm_nb_reduced_clusterer_file, \"salm_nb\", salm_nb_reseeded_dir, salm_nb_reduced_tica_file, salm_nb_traj_files)\n",
    "\n",
    "\n",
    "samples_dir = \"%s/salm_nb_clusterer_%dclusters_%dsamples_samples_kdtree\" %(tica_dir, n_clusters, n_samples)\n",
    "reseed_dir = \"%s/salm_nb_clusterer_%dclusters_%dsamples_samples_kdtree_reseed\" %(tica_dir, n_clusters, n_samples)\n",
    "samples_indices_file = \"%s/salm_nb_clusterer_%dclusters_%dsamples_samples_kdtree_indices.h5\" %(tica_dir, n_clusters, n_samples)\n",
    "if not os.path.exists(samples_dir): os.makedirs(samples_dir)\n",
    "if not os.path.exists(reseed_dir): os.makedirs(reseed_dir)\n",
    "print(samples_dir)\n",
    "traj_files = get_trajectory_files(traj_dir, traj_ext)    \n",
    "sample_from_clusterer(salm_nb_reduced_clusterer_file, salm_nb_reduced_tica_file, [traj_files[i] for i in salm_nb_trajs], \n",
    "                      n_samples, samples_dir, samples_indices_file,\n",
    "                      worker_pool=dview, parallel=False, lig_name=\"LIG\", reseed_dir = reseed_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import custom_clusterer\n",
    "reload(custom_clusterer)\n",
    "from custom_clusterer import *\n",
    "\n",
    "n_samples = 1\n",
    "n_clusters = 300\n",
    "\n",
    "traj_files = get_trajectory_files(traj_dir, traj_ext)\n",
    "featurized_trajectories = [traj_files[i] for i in feature_indices]\n",
    "salm_traj_files = [t for t in featurized_trajectories if \"salm\" in t]\n",
    "reduced_clusterer = compat_verboseload(clusterer300_clusters)\n",
    "salm_reduced_clusterer_file = \"%s/salm_reduced_clusterer.h5\" %tica_dir\n",
    "\n",
    "salm_reduced_tica_file = \"%s/salm_reduced_tica.h5\" %tica_dir\n",
    "salm_reduced_tica = [x[:,range(1,5)] for x in salm_tica]\n",
    "salm_reseeded_dir = \"%s/salm_reseeded_samples\" %tica_dir\n",
    "if not os.path.exists(salm_reseeded_dir):\n",
    "    os.makedirs(salm_reseeded_dir)\n",
    "verbosedump(salm_reduced_tica, salm_reduced_tica_file)\n",
    "\n",
    "salm_reduced_clusters = [reduced_clusterer.labels_[i] for i in salm_trajs]\n",
    "salm_reduced_clusterer = copy.deepcopy(reduced_clusterer)\n",
    "salm_reduced_clusterer.labels_ = salm_reduced_clusters\n",
    "verbosedump(salm_reduced_clusterer, salm_reduced_clusterer_file)\n",
    "#reseed_from_clusterer(salm_reduced_clusterer_file, \"salm\", salm_reseeded_dir, salm_reduced_tica_file, salm_traj_files)\n",
    "\n",
    "\n",
    "samples_dir = \"%s/salm_clusterer_%dclusters_%dsamples_samples_kdtree\" %(tica_dir, n_clusters, n_samples)\n",
    "reseed_dir = \"%s/salm_clusterer_%dclusters_%dsamples_samples_kdtree_reseed\" %(tica_dir, n_clusters, n_samples)\n",
    "samples_indices_file = \"%s/salm_clusterer_%dclusters_%dsamples_samples_kdtree_indices.h5\" %(tica_dir, n_clusters, n_samples)\n",
    "if not os.path.exists(samples_dir): os.makedirs(samples_dir)\n",
    "if not os.path.exists(reseed_dir): os.makedirs(reseed_dir)\n",
    "print(samples_dir)\n",
    "traj_files = get_trajectory_files(traj_dir, traj_ext)    \n",
    "sample_from_clusterer(salm_reduced_clusterer_file, salm_reduced_tica_file, [traj_files[i] for i in salm_trajs], \n",
    "                      n_samples, samples_dir, samples_indices_file,\n",
    "                      worker_pool=dview, parallel=False, lig_name=\"LIG\", reseed_dir = reseed_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import custom_clusterer\n",
    "reload(custom_clusterer)\n",
    "from custom_clusterer import *\n",
    "\n",
    "n_samples = 1\n",
    "n_clusters = 300\n",
    "\n",
    "traj_files = get_trajectory_files(traj_dir, traj_ext)\n",
    "featurized_trajectories = [traj_files[i] for i in feature_indices]\n",
    "apo_traj_files = [t for t in featurized_trajectories if \"apo\" in t]\n",
    "reduced_clusterer = compat_verboseload(clusterer300_clusters)\n",
    "apo_reduced_clusterer_file = \"%s/apo_reduced_clusterer.h5\" %tica_dir\n",
    "\n",
    "apo_reduced_tica_file = \"%s/apo_reduced_tica.h5\" %tica_dir\n",
    "apo_reduced_tica = [x[:,range(1,5)] for x in apo_tica]\n",
    "apo_reseeded_dir = \"%s/apo_reseeded_samples\" %tica_dir\n",
    "if not os.path.exists(apo_reseeded_dir):\n",
    "    os.makedirs(apo_reseeded_dir)\n",
    "verbosedump(apo_reduced_tica, apo_reduced_tica_file)\n",
    "\n",
    "apo_reduced_clusters = [reduced_clusterer.labels_[i] for i in apo_trajs]\n",
    "apo_reduced_clusterer = copy.deepcopy(reduced_clusterer)\n",
    "apo_reduced_clusterer.labels_ = apo_reduced_clusters\n",
    "verbosedump(apo_reduced_clusterer, apo_reduced_clusterer_file)\n",
    "#reseed_from_clusterer(apo_reduced_clusterer_file, \"apo\", apo_reseeded_dir, apo_reduced_tica_file, apo_traj_files)\n",
    "\n",
    "\n",
    "samples_dir = \"%s/apo_clusterer_%dclusters_%dsamples_samples_kdtree\" %(tica_dir, n_clusters, n_samples)\n",
    "reseed_dir = \"%s/apo_clusterer_%dclusters_%dsamples_samples_kdtree_reseed\" %(tica_dir, n_clusters, n_samples)\n",
    "samples_indices_file = \"%s/apo_clusterer_%dclusters_%dsamples_samples_kdtree_indices.h5\" %(tica_dir, n_clusters, n_samples)\n",
    "if not os.path.exists(samples_dir): os.makedirs(samples_dir)\n",
    "if not os.path.exists(reseed_dir): os.makedirs(reseed_dir)\n",
    "print(samples_dir)\n",
    "traj_files = get_trajectory_files(traj_dir, traj_ext)    \n",
    "sample_from_clusterer(apo_reduced_clusterer_file, apo_reduced_tica_file, [traj_files[i] for i in apo_trajs], \n",
    "                      n_samples, samples_dir, samples_indices_file,\n",
    "                      worker_pool=dview, parallel=False, lig_name=\"LIG\", reseed_dir = reseed_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import custom_clusterer\n",
    "reload(custom_clusterer)\n",
    "from custom_clusterer import *\n",
    "traj_files = get_trajectory_files(traj_dir, traj_ext)\n",
    "featurized_trajectories = [traj_files[i] for i in feature_indices]\n",
    "salm_nb_traj_files = [t for t in featurized_trajectories if \"salm_nb\" in t]\n",
    "reduced_clusterer = compat_verboseload(clusterer300_clusters)\n",
    "salm_nb_reduced_clusterer_file = \"%s/salm_nb_reduced_clusterer.h5\" %tica_dir\n",
    "\n",
    "salm_nb_reduced_tica_file = \"%s/salm_nb_reduced_tica.h5\" %tica_dir\n",
    "salm_nb_reduced_tica = [x[:,range(1,10)] for x in salm_nb_tica]\n",
    "verbosedump(salm_nb_reduced_tica, salm_nb_reduced_tica_file)\n",
    "\n",
    "salm_nb_reduced_clusters = [reduced_clusterer.labels_[i] for i in salm_nb_trajs]\n",
    "salm_nb_reduced_clusterer = copy.deepcopy(reduced_clusterer)\n",
    "salm_nb_reduced_clusterer.labels_ = salm_nb_reduced_clusters\n",
    "verbosedump(salm_nb_reduced_clusterer, salm_nb_reduced_clusterer_file)\n",
    "reseed_from_clusterer(salm_nb_reduced_clusterer_file, \"salm_nb\", tica_dir, salm_nb_reduced_tica_file, salm_nb_traj_files)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "closest_indices = find_closest_indices_to_cluster_center(tics_1_3_file, clusterer16_clusters)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "closest_indices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from scipy.stats import gamma \n",
    "x = np.linspace(0.,16.0, 100)\n",
    "y = gamma.pdf(x, 5.0, loc=8.0, scale=1)\n",
    "plt.plot(x,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from scipy.stats import gamma \n",
    "x = np.linspace(0.,16.0, 100)\n",
    "y = gamma.pdf(x, 3.0, loc=8.0, scale=0.4)\n",
    "plt.plot(x,y)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "reseed_from_clusterer(salm_nb_clusterer_file, \"salm_nb\", tica_dir, salm_nb_tica_file, [trajs[i] for i in salm_nb_trajs])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "min_pop = [3, 4, 6, 8, 9, 26, 29, 30, 31, 32, 33, 39, 41, 43, 52, 55]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "plot_all_tics_and_clusters(tica_dir, salm_nb_tica_file, salm_nb_clusterer_file, lag_time, tic_range=range(0,5), main = \"salm_nb_Reseeds\", label = \"cluster_id\", active_cluster_ids = min_pop, intermediate_cluster_ids = [], inactive_cluster_ids = [], concatenate=True, axes=tica_axes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}