search

Rose-STL-Lab / LIMO

generative model for drug discovery
59 stars 14 forks source link

Generating new binding_affinity.pt files #1

Closed slochower closed 11 months ago

slochower commented 2 years ago

Hi. In the README, you write:

The default binding affinity model is ESR1 (which is binding_affinity.pt), but to optimize binding to another protein one must make sure the binding_affinity.pt file contains the model for that correct protein.

Can you describe a bit more how you generate the file binding_affinity.pt for a new protein? Thanks!

PeterEckmann1 commented 2 years ago

Hi, thanks for your interest! binding_affinity.ptis a pytorch model file, so it will require retraining of the property predictor (see https://github.com/Rose-STL-Lab/LIMO#training-the-property-predictor).

If you have a new protein, first you need to prepare files for docking with AutoDock-GPU. Instructions for that can be found at: https://autodock.scripps.edu/wp-content/uploads/sites/56/2021/10/AutoDock4.2.6_UserGuide.pdf starting at page 13. Once you have all the files that are needed (it should look similar to the files in https://github.com/Rose-STL-Lab/LIMO/tree/main/1err), put them in a folder in the main directory and then call python train_property_predictor.py --prop binding_affinity --protein_file <folder name>/<protein name>.maps.fld. When the training is done, a new binding_affinity.pt file will be saved, and you should be ready to generate molecules!

Let me know if you need any assistance during this process, I'm happy to help :)