cvigilv / simspread

De novo target prediction by chemical similarity-guided network-based inference

De novo prediction of drug targets and candidates by chemical similarity-guided network-based inference

This repo contains the scripts for reproducing the results showcased in Vigil, Schuller (2022) "De novo prediction of drug targets and candidates by chemical similarity-guided network-based inference".

Table of Contents

• Requirements
• Repository description
• Usage
• Contact

Requirements

Python requirements:

• python >= 3.9
• matplotlib >= 3.5.1
• seaborn>= 0.11.2
• scikit-learn>= 1.0.2
• pandas>= 1.4.1
• tqdm >= 4.62.3

Julia requirements:

• julia >= 1.7.2
• CUDA.jl >= 3.8.5
• ArgParse.jl >= 1.1.4
• NamedArrays.jl >= 0.9.6

Other:

• bash
• jupyter-notebook

Repository description

This repository has the following organization:

.
├── bin                 # Scripts to run predictions
│  └── predict
│     ├── 10fold
│     ├── loo
│     └── timesplit
├── data                # Datasets used in study
│  ├── chembl
│  ├── wu2017
│  └── yamanishi2008
├── results             # Results obtained
│  ├── 10fold
│  ├── 10fold_dti
│  ├── loo
│  └── timesplit
└── src                 # Scripts needed to run predictions
   ├── evaluate
   ├── modules
   └── predict
      ├── 10fold
      ├── loo
      └── timesplit

For each directory, a corresponding README is available for further information

Usage

1. Clone this repo (for help see this tutorial).
2. Scripts used to generate predictions are kept here.
3. Scripts needed for predictions are kept here.
4. Datasets used in study are kept here.

Contact

Any question, suggestion, advice and/or help needed to reproduce results, please contact Carlos Vigil Vásquez @ [e-mail].