Integration of persistent Laplacian and pre-trained transformer for protein solubility changes upon mutation

This code is for Integration of persistent Laplacian and pre-trained transformer for protein solubility changes upon mutation.

---

Python Environment needed

• fair-esm 2.0.0
• numpy 1.23.5
• scipy 1.11.3
• scikit-learn 1.3.2
• python 3.10.12
• Biopython 1.80
• Softwares to be installed for ./bin folder (See README at bin)

File description

There are five folders after extracted.

• bin
  • jackal.dir
  • jackal_64bit
  • PPIprepare
  • PPIstructure
  • profix
  • scap
  • SPIDER2_local
  • mibpb5
  • MS_Intersection
• dataset
  • alphafold_pdb.zip (at root directory)
• mutsol
  • mutsol.csv
  • mutsol.npy
• model and feature generation
  • mutsol_build_features.py
  • mutsol_build_cv.py
  • mutsol_blindtest.py
  • mutsol_10cv.py

Feature generation

The dependencies libraries (the current version is only test on linux system):

• BLAST+/2.10.1
• dssp/3.1.4
• vmd (should be a global call)
• pdb2pqr (should be a global call)
• mibpb5 (should be a globale call, which can be used from https://weilab.math.msu.edu/MIBPB/)