search

dauparas / LigandMPNN

MIT License
238 stars 47 forks source link

LigandMPNN dataset #3

Closed bio-rat closed 9 months ago

bio-rat commented 10 months ago

Hi,

Thank you for creating this awesome program. I have been using this on a crystal structure currently on PDB. I just want to know if it was in the training dataset. Where do I look up if my protein is in the dataset or not?

Have a good day!

dauparas commented 10 months ago

Hey!

Your PDB is likely in our training set if deposited before Dec 16, 2022, has a lower than 3.5 A resolution, and has a total number of residues in the PDB biounit smaller than 6000 residues. You can find details on the training data in the preprint (https://www.biorxiv.org/content/10.1101/2023.12.22.573103v1.full.pdf).