jrem-chem / gpcrlignet

Code for an artificial neural net classifier of small molecule GPCR activity

GPCRLigNet

Contains python scripts described by the methods in the paper "GPCRLigNet: Rapid Screening for GPCR Active Ligands Using Machine Learning" https://link.springer.com/article/10.1007/s10822-023-00497-2

How to make predictions with GPCRLigNet

1. edit the following file to your needs for generating the morgan fingerprints from smiles strings: druglikeness_analysis/generate_fpts.py
2. edit the following file to turn those fingerprints into logit GPCR activity predictions: druglikeness_analysis/make_predictions.py
3. the best performing model was models/cicular_4_models_6_17_21/model_cicular4.tf
4. other models can be found in models/cicular_4_models_6_17_21/

Unfortunately the full datasets were too large for github, if you would like to see anything else please dont hesitate to reachout.

gpcrlignet's Issues

request for access to training dataset

Hello ,

I'm currently exploring GPCR and I'm impressed by your project, GPCRLigNet. Could you kindly share the training data used for this project? Understanding the data would greatly benefit my work.

If there are any conditions or guidelines for accessing the data, I'm happy to comply.

Thank you for your time and effort.

Best