DeepImmuno

PaperCodeProduction Ready

Prediction of immunogenic epitopes for T cell immunity (Class 1)

Readiness

Readiness: Production Ready
Smoke: passed
Artifacts: passed
Production: production_pass
Storage: <5
Strategy: shared deepfri_tf23.sif TensorFlow runtime + DeepImmuno py36 pandas target layer
Next: Production Ready. Optional hardening: add UI fields for peptide/HLA batch input, support uploaded CSVs, and expose score-threshold filtering.

TensorFlow 2.3 environment/container unavailable in first pass. 2026-07-08T23:03:57+08:00 SOP production validation: promoted from Production Blocked to Production Ready. Materialized DeepImmuno Python 3.6 pandas layer at /media/nik/seagate_nik/bio_server/tools/deepimmuno/py36_site and reused deepfri_tf23.sif plus deepfri_tf23_py36_site. Validated upstream frankligy/DeepImmuno deepimmuno-cnn.py --mode multiple through Ubuntu runner API job adbf97cef1fb / Slurm job 573. Run completed in 3.674 s on bundled peptide/HLA benchmark and produced 3 finite immunogenicity scores; max 0.8267597 for CCCCCCCCC/HLA-B*5801. Artifacts verified: summary.json, results.csv, deepimmuno-cnn-result.txt, deepimmuno_input.csv, deepimmuno_stdout.txt, deepimmuno_stderr.txt, slurm.out, slurm.err.