AF3 WebServer AlphaPullDown automation

Generate reusable JSON job bundles from BioGRID interactor exports and seed-sequence bait protein inputs.

What this repo requires

• A config file that lists your bait proteins and job-set definitions.
• An input directory containing:
  • main_sequences.txt with named protein sequence blocks.
  • One BioGRID-style export per seed protein, named <PROTEIN>.txt.
• Optionally, a cached accession-to-sequence TSV.

Quick start

1. Edit configs/job_sets.json to match your proteins and grouping strategy.
2. Put your input files under input_data/.
3. Run:

python generate_outputs.py --repo-root .

Generated files are written to generated_job_output/ by default.

Input format

main_sequences.txt uses named blocks like:

PROTEIN_A:
MSEQUENCEEXAMPLE

PROTEIN_B:
MSEQUENCEEXAMPLE

Each per-protein interaction file should be a BioGRID-style tab3 txt export for that protein.

Notes

• configs/job_sets.json is only a generic example; replace every placeholder before use.