AlexanderKroll / ESP

MIT License
68 stars 23 forks source link

Description

This repository contains the code and datasets to reproduce the results and figures and to train the models from our paper "The substrate scopes of enzymes: a general prediction model based on machine and deep learning".

For people interested in using the trained prediction model, we implemented a web server that allows an easy use of our trained model. The prediction tool can be run in a web-browser and does not require the installation of any software. Prediction results are usually ready within a few minutes. Example inputs can be found on the homepage.

For people interested in using a python function to achieve predictions of the trained model, we created a GitHub repository that allows an easy use of our trained model.

Downloading data folder

Before you can run all scripts of this repository, you need to download and unzip an additional data folder from Zenodo. Afterwards, this repository should have the following strcuture:

├── notebooks_and_code
├── data
├── additional_data_ESP            
└── README.md

Using code and reporducing results

All code to reproduce the results is available in the form of Jupyter Notebooks in the folder "notebooks_and_code". All code and produced output files are available in the folder "data".

Requirements for running the code in this GitHub repository

The code was implemented and tested on Windows with the following packages and versions (installation took ~20 minutes)

The listed packaged can be installed using conda and pip:

pip install torch
pip install numpy
pip install tensorflow
pip install fair-esm
pip install jupyter
pip install matplotlib
pip install hyperopt
pip install pickle
pip install biopython
conda install pandas=1.3.0
conda install -c conda-forge py-xgboost=1.2.0
conda install -c rdkit rdkit