search

zhenglz / onionnet-v2

An updated version of OnionNet
GNU General Public License v3.0
4 stars 2 forks source link

Availability of models #1

Open hypnopump opened 4 years ago

hypnopump commented 4 years ago

I've checked the V1 and it seems a good take. Are the models available for this V2? Is this project in active development? If so, how could I help?

zhenglz commented 4 years ago

Here in v2, I was planning considering protein-ligand per-residue contacts and protein shape (C-alpha contact map) as additional features, but I am not sure whether the additional features will be useful for binding affinity prediction.

Currently, this project is inactive, and the features have not been finalized, the deep learning model has not been trained. But if you are interested in developing good binding affinity models, we could work on it together. I would share dataset and discuss and develope ideas together.

My plan would be, firstly based on v1, additional protein-ligand per-residue based features are designed, and then trained. Finally, the model could be tested on docking enrichment experiments (such as DUDE datasets), or docking-pose rescoring experiments.

hypnopump commented 4 years ago

Hmmm seems nice. I'm very interested in good affinity models. I've tried v1 and it's nice for values lower than 8, but it stops increasing past there. It's almost impossible to get a value bigger than 9.5. I think the main improvements to v1 might come for extending the range of applicability. I've got an idea that could work. Is the dataset used to train the v1 available somewhere?

Also, let's follow with this conversation via email: ericalcaide1@gmail.com