MaSIF

5-30 minutesPaperCodeNeed >20GB

Molecular surface interaction fingerprints.

Readiness

Readiness: Need >20GB
Smoke: not_tested
Artifacts: not_applicable
Production: need_gt20gb_legacy_tf_surface_stack
Storage: >20
Strategy: legacy MaSIF container plus large surface/precompute data phase
Next: Defer to >20GB legacy-container/data round; build dedicated MaSIF container, stage minimal pretrained/example data, then run MaSIF-site or ligand Slurm benchmark before adapter wiring.

MaSIF source cloned from LPDI-EPFL/masif rev 2a370518e0d0d0b0d6f153f2f10f6630ae91f149; source footprint ~843 MB. README requires Python 3.6, TensorFlow 1.9, PyMesh, reduce, MSMS, PDB2PQR/APBS/multivalue/open3D. Documented full data can consume ~1.4 TB, and MaSIF-ligand precompute alone notes ~400 GB. No benchmark or API adapter run. 2026-07-05 SOP triage: readiness_status already indicated need_gt20gb_storage while label was stale Smoke Pending. MaSIF requires legacy Python 3.6/TensorFlow 1.9/PyMesh/reduce/MSMS/PDB2PQR/APBS/open3D stack; documented full data can reach about 1.4 TB and MaSIF-ligand precompute alone about 400 GB. Reclassified to Need >20GB.