File size: 4,979 Bytes
f12d3b3 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 | # Model artifacts
Every local artifact the PROTAC pipeline loads. Run `python scripts/verify_release.py`
to check byte counts, SHA-256 hashes and private-path hygiene. The machine-readable
manifest is `docs/artifacts.json`.
| Artifact | Purpose | Size | SHA-256 |
|---|---|---:|---|
| `checkpoints/protac_linker_finetune/epoch=99-v1.ckpt` | DeFoG generator, fine-tuned on PROTAC-DB linkers | 255,211,895 B | `8cfb48d9f5b3bcabb28c4dce7548e1a792494c48f05150a1b262b5215c1da78f` |
| `src/classifier/protac_linker_clf.pkl` | linker classifier (generic) | 128,117 B | `fa667025962d132ce9f2ebf8f2dc49752416ed154792ca0c87344335fa714462` |
| `src/classifier/protac_linker_clf_BRD9VHL.pkl` | linker classifier (BRD9VHL) | 28,117 B | `eae385be1576fec674b8a537e50b56e1ab1fc47599a82ae277912d958c0b7d6f` |
| `src/classifier/protac_linker_clf_CDK6CRBN.pkl` | linker classifier (CDK6CRBN) | 30,035 B | `85bd40599d82c5cc8a35cc844d12d1066a0d4699bcc7e229937e1fe1d3daad3d` |
| `src/classifier/protac_linker_clf_ErlotinibVHL.pkl` | linker classifier (ErlotinibVHL) | 26,397 B | `e031d310829c7d196887e2fd077eff95be445ed1984516540ac69308c26384ee` |
| `src/classifier/protac_linker_clf_amine.pkl` | linker classifier (amine) | 76,524 B | `e0e851f41a004ea26e3b66caef83bf6ab4fa2229447427ab996e8a195c0df2d2` |
| `src/classifier/protac_linker_clf_amino_acid.pkl` | linker classifier (amino_acid) | 188,080 B | `2558addea35da16ddf82affaf68ea9046bc5b9b3d307f6ee668c117df2798428` |
| `src/degrademaster/data/case_study/arcrbn_pocket_graph.pt` | ARCRBN pocket graph | 35,405 B | `98f70b82b731760b2b25b9ae0e7b1a917a88fd215f2ba4e61cf1654207d137e3` |
| `src/degrademaster/data/case_study/brd4_pocket_graph.pt` | BRD4 pocket graph | 15,675 B | `8dbda805c329bc5804d95983acaf4260e712bfa8fefcc5f39d76b4d2cd3583b9` |
| `src/degrademaster/data/case_study/brd9_pocket_graph.pt` | BRD9 pocket graph | 201,019 B | `89346178dfdc8fd277c57d08293a903133a263b1206184224cf250f71ad8f3a0` |
| `src/degrademaster/data/case_study/cdk6crbn_pocket_graph.pt` | CDK6CRBN pocket graph | 32,607 B | `0193874c2150e40a2e8c7a6b4eddf1c9b1768ac0573716b0fbf79e365de62310` |
| `src/degrademaster/data/case_study/crbn_pocket_graph.pt` | CRBN pocket graph | 25,467 B | `6e4e8e6756a343379118150370061f1b781d36f1bc681693d6d7fd8a8d0bce12` |
| `src/degrademaster/data/case_study/egfr_pocket_graph.pt` | EGFR pocket graph | 32,602 B | `5d86cd0d80363bab3c2835aefa817181cd96b2e0c14b35c55e384cf8e9dc64c1` |
| `src/degrademaster/data/case_study/krasvhl_pocket_graph.pt` | KRASVHL pocket graph | 25,686 B | `79efc62c546142277c6cba995d7926ca0dc3b600c2e98cce766519054243284c` |
| `src/degrademaster/data/case_study/processed/ligase_pocket.pt` | VHL E3 pocket graph (collated) | 419,524 B | `3b35bb97bddce48ccc5ebddbdb435a98654b94f569441cf6f8bc66e3a11e4d38` |
| `src/degrademaster/data/case_study/arcrbn_target_feature.npy` | 30-d feature vector | 368 B | `9ef6c2c04f54c6585cdac4e3d0ffb38c15ba7d1912a29f25f6a4967c7e13799e` |
| `src/degrademaster/data/case_study/brd4_target_feature.npy` | 30-d feature vector | 368 B | `28a1903da353ff6a92c4d4b8828e66e7264bcd4f07a27329d2f898cf000f63e4` |
| `src/degrademaster/data/case_study/brd4crbn_target_feature.npy` | 30-d feature vector | 368 B | `28a1903da353ff6a92c4d4b8828e66e7264bcd4f07a27329d2f898cf000f63e4` |
| `src/degrademaster/data/case_study/brd9_target_feature.npy` | 30-d feature vector | 368 B | `dbe5d7e51907b44396a891a574010544f9e7461c09618d940fe26ddf59ef8473` |
| `src/degrademaster/data/case_study/cdk6crbn_target_feature.npy` | 30-d feature vector | 368 B | `4da1adf39d018d61cd80a67f5950eaf734b425b69690a157c392d792b020712a` |
| `src/degrademaster/data/case_study/crbn_e3_feature.npy` | 30-d feature vector | 368 B | `71f830577fd901b6be42b9bb8f02f12e25245aefafac709d0b68910dfa8214db` |
| `src/degrademaster/data/case_study/egfr_target_feature.npy` | 30-d feature vector | 368 B | `c31f02bfeb6ce63b0fcf41a7b235556dc2b76ab6054626116c08434a29b639de` |
| `src/degrademaster/data/case_study/krasvhl_target_feature.npy` | 30-d feature vector | 368 B | `51d50af89f905b6a4eb1f1d2b90d9ca8a2ee7e7dbace16fdab56e159f2c0385f` |
| `src/degrademaster/data/case_study/vhl_e3_feature_egfr.npy` | 30-d feature vector | 368 B | `aca7d4ba3f95a2d903e542f761c3d448323d7ff8a9e1d58e58d9b8bd55d0a74a` |
| `src/degrademaster/data/PROTAC/name.json` | PROTAC-DB entry table (property windows, similarity) | 3,419,093 B | `bbe5ffe734f416ecf153f4ab42a17c6a3a86a79cfcb0faa5a5f13a672733d575` |
| `data/linker.csv` | PROTAC-DB derived linker table | 703,569 B | `73e4be1680fe9e2e2cf7256ff3fa9423a9a7e480d9b1baa529081a2bc89ac5b1` |
DegradeMaster and EGFR-PROPK are published upstream without a license and are therefore
not redistributed; `python scripts/fetch_third_party.py` retrieves them at the commits
used for the paper. Their absence is reported, not failed, by the verifier.
The ORCA graphlet-counting C++ sources bundled with upstream DeFoG are excluded: they
carry no license header and are unused by the PROTAC pipeline.
|