PLAbDab

1-5 minutesPaperCodeNeed >5GB

Search antibody sequences from patents and literature

Readiness

Readiness: Need >5GB
Smoke: not_tested
Artifacts: not_applicable
Production: need_gt5gb_database_and_overlay_pending
Storage: <5
Strategy: compiled/binary SIF or shared python312_slim + ext3 overlay
Next: Build PLAbDab overlay on a compatible antibody runtime, download database once to shared storage, then run Slurm search benchmark before wiring.

PLAbDab source staged from oxpig/PLAbDab rev 6820b1382a91c8354f20e96d4bb531d46008f13d; license BSD-3-Clause. Source footprint ~28 MB. Data archive plabdab_data.tar.gz verified by HEAD only: 5,045,787,915 bytes compressed. Login-node import and Slurm diagnostic job 401 failed due missing pandas/numba/torch/kasearch; existing ImmuneBuilder runtime is a reuse candidate. No functional search smoke. 2026-07-04 backend update: PLAbDab data archive is about 5.0GB compressed and current shared SIFs lack pandas/numba/torch/kasearch, so this belongs in the >5GB database/overlay phase.