General Query - PPIFold Setup & Documentation

[unikill066]

Could someone check that my PPIFold setup is correct—Alphafold, SignalP, Singularity (with the image) installed, the project folder laid out like this:

PPIFold_test/
  ├─ alphafold/
  ├─ test.txt
  ├─ feature/
  ├─ alpha-analysis_jax_0.4.sif
  ├─ conf.txt
  ├─ test.fasta
  └─ PPI.log

and conf.txt set to

Path_Uniprot_ID       : /mnt/WorkingDos/ppifold_testing/test.txt
Path_AlphaFold_Data   : /mnt/WorkingDos/ppifold_testing/alphafold
Path_Singularity_Image: /mnt/WorkingDos/ppifold_testing/alpha-analysis_jax_0.4.sif
Path_Pickle_Feature   : /mnt/WorkingDos/ppifold_testing/feature

—then point me to documentation for actually running PPIFold and, ideally, show how I might script a workflow that screens one anchor gene (e.g., SNAP25) against a list of genes/sequences, please?

[Qrouger]

Hi @unikill066, thanks for your interest in our pipeline !
Your directory structure looks good. But your conf file should point directly to your folder, like : /mnt/WorkingDos/PPIFold_test/test.txt
PPIFold is a tool designed to naively generate an interactome and homo-oligomerization from a list of proteins. It's not coded and optimized for make bait versus prey like you want.
That said, you can still use it that way: simply place your bait protein as the first entry in test.txt. The pipeline will start by comparing this bait to all the others. You can then stop the run once these comparisons are completed.

To run PPIFold you need to activate your conda environnement and launch this command in your folder : PPIFold --use_mmseq Boolean --make_multimers Boolean --max_aa Integer --use_signalP Boolean --org String
Argument are optional. See documentation here : https://github.com/Qrouger/PPIFold?tab=readme-ov-file#arguments

Finally, your test.txt should contain a comma-separated list of UniProt IDs or sequences, like:

P60880, UniprotID2, UniprotID3 ...

Best,
Q.