search

baker-laboratory / rf_diffusion_all_atom

Public RFDiffusionAA repo
Other
321 stars 50 forks source link

RFdiff_aa for covalent small molecule binder and additional residues #13

Open Xavier-BZT opened 7 months ago

Xavier-BZT commented 7 months ago

Hi,

I am testing RFDaa for the design of small molecule binders where the small molecules are covalently bound to the protein, a bit similar to the bilin reported in the paper.

I would like to keep some specific residues (the one covalently bound) but also some distal residues.

What would be the right syntax to use in the contigmap.contigs ?

Many thanks.

Xavier-BZT commented 7 months ago

I think that I've found the solution with the ppi.hotspot_res=[Ax] card.

Xavier-BZT commented 6 months ago

I still have some doubts about the ppi_hotspot_res. Is it only for protein-protein-interaction or is it also ok for protein-ligand interaction ? Other instructions could be more specific in such case. Any help would be much appreciated.

dohyeonscottkim commented 4 months ago

You can execute the command as follows:

python run_inference.py inference.deterministic=True diffuser.T=100 inference.output_prefix=output/ligand_protein_motif/sample inference.input_pdb=input/1haz.pdb contigmap.contigs="['10-120,A84-87,20-40,A117-121,10-100']" inference.ligand=CYC inference.num_designs=1 inference.design_startnum=0

It seems there is a difference in argument parsing when using apptainer compared to running the script locally. Specifically, the contigmap.contigs argument should use double quotes around the entire value and single quotes inside the brackets for the ranges, like this: contigmap.contigs="['10-120,A84-87,20-40,A117-121,10-100']".