Deep learning-driven prediction of drug mechanism of action from large-scale chemical-genetic interaction profiles
This repository provides codes and materials to accompany the paper "Deep learning-driven prediction of drug mechanism of action from large-scale chemical-genetic interaction profiles"
Framework
├── Codes/
│ ├── Gene_clustering/
│ ├── Hierarchical_clustering/
│ └── hierarchical_clustering.R
│ └── Homology_search/
│ ├── Ecoli_homologs_Rproj_020921.Rproj
│ └── Mtub_Ecoli_homologs_gene.Rmd
│ └── Mtb_pred_ML/
│ ├── main.ipynb
│ └── data_processing.ipynb
├── Data/
│ ├── BLAST/
│ ├── Ecoli_K12_MG1655_GCF_000005845_2_ASM584v2_genomic.gff
│ ├── Ecoli_K12_MG1655_GCF_000005845_2_ASM584v2_protein.faa
│ ├── Ecoli_K12_MG1655_GCF_000005845_2_ASM584v2_protein.faa.phr
│ ├── Ecoli_K12_MG1655_GCF_000005845_2_ASM584v2_protein.faa.pin
│ ├── Ecoli_K12_MG1655_GCF_000005845_2_ASM584v2_protein.faa.psq
│ ├── Mtb_genes_of interest_names.txt
│ ├── Mtuberculosis_H37RvGCF_000195955_2_ASM19595v2_genomic.gff
│ └── Mtuberculosis_H37RvGCF_000195955_2_ASM19595v2_protein.faa
│ ├── CGIP/
│ └── 47K_SMILES_Zscore.csv
│ ├── Johnson_data/
│ └── README.md
│ └── ML_data/
│ ├── Data_split/
│ ├── Features/
│ ├── MB_test.npz
│ ├── MB_train.npz
│ ├── MB_val.npz
│ ├── RN_test.npz
│ ├── RN_train.npz
│ └── RN_val.npz
│ ├── Opt_hyperpars/
│ ├── DMPNN_RN.json
│ ├── FFN_MB.json
│ ├── FFN_RN.json
│ └── MPNN_RN.json
│ ├── test.csv
│ ├── train.csv
│ └── val.csv
│ └── binary_13_clusters.csv
├── Image/
│ └── framework.png
├── Results/
│ ├── Homology_results/
│ ├── Ecoli_vs_Mtuberculosis_outfmt7.tsv
│ ├── Mtb_to_Ecoli_154_Ecoli_gff.txt
│ └── Mtub_to_Ecoli_154_Ecoli_gff.txt
│ ├── Mtb_inhibitors_pred/
│ ├── Mtb_inhibitors_DMPNN_preds.csv
│ ├── Mtb_inhibitors_FFN_MB_preds.csv
│ ├── Mtb_inhibitors_FFN_RN_preds.csv
│ └── Mtb_inhibitors_MPNN_preds.csv
│ └── Trained_model
│ ├── DMPNN_RN_Ensemble_5/
│ └── ......
│ ├── FFN_MB_Ensemble_5/
│ └── ......
│ ├── FFN_MB_Ensemble_5/
│ └── ......
│ └── MPNN_RN_Ensemble_5/
│ └── ......
├── README.md
This repository contains scripts, datasets and partial results that support the findings of our study. The Johnson et al. data investigated in this study is publicly accessible on the web site. The code to train the Message Passing Neural Networks is available on Chemprop.