TCRModel2

PaperCodeNeed >20GB

Predict T Cell Receptor structures

Readiness

Readiness: Need >20GB
Smoke: passed
Artifacts: passed
Production: deferred_large_assets
Storage: >20
Strategy: shared python312_slim.sif + tool ext3 overlay
Next: Stage shared AlphaFold/TCRModel2 databases and CUDA/Singularity image, then run upstream example benchmark before wiring runner.

Storage correction 2026-07-01: upstream TCRModel2 requires AlphaFold-derived assets including PDB mmCIF and params estimated around 120 GB after unzip plus CUDA/GPU runtime. Upstream source https://github.com/piercelab/tcrmodel2 revision 90536deaf766fe169f51679ce9456f3b165be5c9. Production validation deferred to >20GB AlphaFold shared asset round; no Slurm scientific benchmark yet.