search

radical-collaboration / IMPRESS

Integrated Machine-learning for PRotEin Structures at Scale
https://radical-project.github.io/impress
1 stars 0 forks source link

Create IMPRESS env on Delta (NCSA) #12

Closed mtitov closed 1 month ago

mtitov commented 1 month ago

Build conda env and test corresponding packages individually

mturilli commented 1 month ago

Document the process and create a dependencies file for RCT. Add this to RP documentation if needed.

mtitov commented 1 month ago

@JonathanEAsh can you please provide (1) test example for PyRosseta, and (2) instructions for ProteinMPNN installation and its corresponding test example.

Related PR https://github.com/radical-collaboration/IMPRESS/pull/14

JonathanEAsh commented 1 month ago

1 - For pyrosetta, I would recommend simply loading a pose and trying to relax it. Make sure you have joey_utils in your working directory. from pyrosetta import * init() from joey_utils import fast_relax_mover pose=pose_from_pdb("FILEPATH") fr=fast_relax_mover() fr.apply(pose) pose.dump_pdb("OUTPATH") 2 - For ProteinMPNN, please follow the instructions here for the environment: https://github.com/dauparas/ProteinMPNN They also have some test examples, with all of the requisite input files already included. Simply run one of those to verify everything is working