Analysis_for_Synergy_Data_Augmentation

This repository provides the analysis process for the synergy data augmentation. It contains the datasets utilized in the analysis and generates the figures that are presented in the paper to describe the analysis process.

Dependencies

1. Pandas 1.1.3
2. Numpy 1.19.2
3. Scipy 1.7.3
4. Matplotlib 3.5.1
5. Seaborn 0.11.2
6. PIL 9.0.1

Usage

The following steps provide a guide to reproduce the analysis results presented in the paper.

Generation of Figure 1

Similarity measure for cellular responses to drug treatment

1. Run python Correlation_drug_pair.py
   This will calculate the Kendall tau correlation coefficients for the drug pairs.
2. Run python Make_pic50_scatter.py
   
3. Run python Legend.py
   
4. Run python Concat_figures.py
   

Generation of Figure 2

Relation between drug similarity and pharmacological effects
Run python Fig2_Analysis_Thresholds.py

Generation of Figure 3

Heatmap of the fraction of drug pairs with positively correlated pharmacological effects
Run python Fig3_Heatmap_Plot.py

Generation of Figure 4

Optimal DACS threshold for data augmentation

1. Download the dataset DACS_score_between_original_drug_and_similar_drugs.csv from the link https://osf.io/kd9e7/ and put the dataset in Data folder
   
2. Run Pre_process_for_the_plot.py
   
3. Run Fig4_Analysis_Intersection.m in MATLAB

Generation of Figure 5

Distribution plots of the synergy scores in original and augmented datasets

1. Download the dataset AugmentedData.csv from link https://osf.io/kd9e7/ and put the dataset in Data folder
   
2. Run Fig5_Histograms.py