Upload folder using huggingface_hub

README.md

@@ -1,3 +1,131 @@
 ---
-license: apache-2.0
+license: mit
+library_name: proteinmpnn
+tags:
+  - protein-design
+  - proteinmpnn
+  - inverse-folding
+  - immunology
+  - mhc
+  - tcr
+  - data-leakage
+  - out-of-distribution
 ---
+
+# ProteinMPNN-noMHC
+
+A retrained **vanilla ProteinMPNN** (Dauparas et al., 2022) on the standard PDB 2021-08-02
+training set **with every MHC-containing structure removed**. The purpose is a leakage-controlled,
+genuinely out-of-distribution model for studying peptide/epitope design on peptide–MHC (pMHC) and
+TCR–pMHC complexes: this model has **never seen an MHC molecule** (classical class I/II or MHC-like)
+or any peptide presented in an MHC groove during training.
+
+- **Architecture:** vanilla ProteinMPNN — `hidden_dim=128`, 3 encoder + 3 decoder layers,
+  `k_neighbors=48`, `backbone_noise=0.20`. **~1.66 M parameters.**
+- **Released checkpoint:** `proteinmpnn_nomhc.pt` (epoch 94; full training target 200 epochs).
+- **Loads directly** with the official `protein_mpnn_run.py` (`--path_to_model_weights <dir> --model_name proteinmpnn_nomhc`).
+- **Base data:** `https://files.ipd.uw.edu/pub/training_sets/pdb_2021aug02.tar.gz`.
+
+## Intended use
+
+Drop-in replacement for vanilla ProteinMPNN when you need an inverse-folding model that is
+**blind to MHC**. Useful for: testing whether epitope/groove-anchor preferences are *learned from
+general protein structure* vs. *memorized from pMHC structures*; OOD design benchmarks on
+immunological complexes; leakage controls for TCR/pMHC modeling pipelines (e.g. the TCera project).
+
+It is **not** a better general-purpose ProteinMPNN — it is a controlled ablation.
+
+## Data partition
+
+Base training set (`pdb_2021aug02`, curated by Ivan Anishchenko): clustered at 30% sequence identity.
+
+| | chains | clusters |
+|---|---:|---:|
+| Full `list.csv` | 555,720 | 27,789 |
+| **Removed (all MHC)** | **6,076** | 27 fully dropped |
+| **Kept (training set used here)** | **549,644** | 27,762 |
+
+`.pt` files in the extracted set: 869,548.
+
+### How "MHC structures" were defined (1,822 PDB IDs)
+
+The exclusion set is the **union** of three complementary sources, to avoid annotation gaps:
+
+1. **RCSB Pfam annotations** — every entry with an MHC-fold polymer entity:
+   `PF00129` (MHC-I / MHC-like, incl. CD1, MR1), `PF00969` (MHC-II β), `PF00993` (MHC-II α) → 1,352 entries.
+2. **STCRDab** TCR–MHC complexes → 456 entries (adds 160 not Pfam-annotated).
+3. **Sequence-signature scan** over all 555,720 chains (MHC-I `SRPG.GEP[RK].I`,
+   β2-microglobulin `YSRHP[AP]ENGK`, MHC-II `HV[IL]IQAEFY` / `FNGTERVR`) → caught **310 more**
+   structures that RCSB had not Pfam-annotated (e.g. `1cd1`).
+
+Full list: `mhc_excluded_pdb_ids.txt`. Of 1,822, **1,531 were present** in this training set.
+
+### Removal is PDB-level (and why)
+
+Each chain is removed if its **PDB entry contains MHC**, dropping the *entire* structure
+(MHC + β2m + peptide + any crystallization Fab/TCR). This is required because the ProteinMPNN
+loader rebuilds the full biological assembly from any one chain — removing only the MHC chain would
+let the β2m or peptide chain re-load the complex (and the presented peptide) into training.
+
+> ⚠️ We deliberately do **not** use cluster-level removal. MHC structures are routinely crystallized
+> with Fabs/TCRs whose Ig V-domains fall into large mixed antibody+TCR clusters; expanding to those
+> clusters would delete thousands of legitimate non-MHC antibody/TCR chains. PDB-level removal keeps
+> antibodies/TCRs from non-MHC structures intact (e.g. the largest Ig-Vκ cluster retained 6,914 of
+> 7,507 chains; only the ~593 anti-MHC Fabs inside MHC complexes were removed).
+
+### Verification
+
+All 9 PDB structures in the RCSB that present the NY-ESO-1 epitope `SLLMWITQC`
+(incl. the pMHC-only `1s9w`, which had leaked the epitope under an earlier "remove TCR–MHC only"
+partition) are confirmed **absent** from the final `list.csv`.
+
+## Training
+
+- Framework: PyTorch 2.1 (CUDA 11.8); 1× NVIDIA A40 (Savio `savio3_gpu`).
+- Command (official `training.py`):
+  `--num_epochs 200 --save_model_every_n_epochs 10 --reload_data_every_n_epochs 2 --num_neighbors 48 --backbone_noise 0.2`
+  (defaults otherwise: `batch_size=10000` tokens, mixed precision, Noam schedule).
+- Convergence: validation sequence-recovery plateaus at ~0.52 (≈ vanilla ProteinMPNN), confirming the
+  ~1% MHC removal did not perturb general learning.
+
+## Evaluation — epitope design on held-out TCR–pMHC (leakage probe)
+
+Design **only the 9-mer peptide** (chain C), MHC + TCR backbones fixed, on five held-out 1G4
+TCR–pMHC complexes (`2BNR, 2F53, 2F54, 2P5E, 2P5W`; native epitope `SLLMWITQC`). 1,000 designs each
+(5,000 total), `sampling_temp=0.1`, seed 37.
+
+**Result: 0 / 5,000 exact epitope recovery** (both this model and a leakier control).
+The discriminating signal is the buried P5 Trp anchor, at matched training length:
+
+| Pos | Native | Leaky control (saw `1s9w`, e41) | **This model (MHC-blind, e44)** |
+|----:|:------:|------:|------:|
+| 5 | **W** | **0.999** | **0.300** |
+| 2 | L | 0.590 | 0.749 |
+| 7 | T | 0.594 | 0.800 |
+| 1/3/4/8/9 | S/L/M/Q/C | ≤0.23 | ≤0.06 |
+
+**Interpretation.** Conditioned on the MHC-groove backbone, the MHC-blind model still designs
+groove-compatible peptides with the correct HLA-A2 anchor grammar (P2→L, C-terminus→L) — i.e. general
+protein-packing rules generalize to the groove. But its confidence at the most MHC-specific buried
+position (W5) drops from ~100% to ~30% once all MHC is removed, and it never recovers the true
+TCR-selected epitope. The near-perfect W5 of the leaky control was largely **memorization** of the
+`1s9w` pMHC, now quantified and removed.
+
+## Files
+
+- `proteinmpnn_nomhc.pt` — model weights (epoch 94).
+- `mhc_excluded_pdb_ids.txt` — the 1,822 excluded MHC PDB IDs.
+- `filter_list.py` — PDB-level `list.csv` filter.
+- `get_mhc_chains.py` — RCSB polymer-entity Pfam → MHC chain resolver.
+
+## Citation
+
+Built on ProteinMPNN:
+
+```bibtex
+@article{dauparas2022robust,
+  title={Robust deep learning--based protein sequence design using ProteinMPNN},
+  author={Dauparas, J. and Anishchenko, I. and Bennett, N. and others},
+  journal={Science}, volume={378}, number={6615}, pages={49--56}, year={2022}
+}
+```

filter_list.py

@@ -0,0 +1,79 @@
+#!/usr/bin/env python
+"""Filter ProteinMPNN list.csv to drop MHC-TCR complexes.
+
+Removes every chain whose 4-char PDB ID is in the TCR-pMHC exclusion set
+(derived from STCRDab: any TCR structure bound to an MHC / MHC-like
+antigen-presenting molecule).  MHC-only and free-TCR structures are kept.
+
+We must drop ALL chains of an excluded PDB (not just the TCR chains),
+because loader_pdb() rebuilds the full biological assembly from any one
+chain, which would re-introduce the TCR-pMHC interface.
+"""
+import argparse, csv, os, shutil
+
+ap = argparse.ArgumentParser()
+ap.add_argument("--data", required=True, help="pdb_2021aug02 dir containing list.csv")
+ap.add_argument("--exclude", required=True, help="text file of PDB IDs to exclude (one per line)")
+args = ap.parse_args()
+
+list_csv = os.path.join(args.data, "list.csv")
+backup = os.path.join(args.data, "list.csv.full")
+
+exclude = set()
+with open(args.exclude) as f:
+    for line in f:
+        p = line.strip().lower()
+        if p:
+            exclude.add(p)
+print(f"exclusion set size: {len(exclude)} PDB IDs")
+
+if not os.path.exists(backup):
+    shutil.copy2(list_csv, backup)
+    print(f"backed up original -> {backup}")
+else:
+    print(f"backup already exists ({backup}); filtering from it")
+
+rows_in = 0
+kept = []
+removed_chains = 0
+removed_pdbs = set()
+matched_excl = set()
+header = None
+with open(backup, newline="") as f:
+    reader = csv.reader(f)
+    header = next(reader)
+    for r in reader:
+        rows_in += 1
+        chainid = r[0]                       # PDBID_CHAINID
+        pdbid = chainid.split("_")[0].lower()
+        if pdbid in exclude:
+            removed_chains += 1
+            removed_pdbs.add(pdbid)
+            matched_excl.add(pdbid)
+        else:
+            kept.append(r)
+
+with open(list_csv, "w", newline="") as f:
+    w = csv.writer(f)
+    w.writerow(header)
+    w.writerows(kept)
+
+clusters_before = set(r[4] for r in kept) if False else None
+# cluster stats from full + filtered
+def clusters_of(path):
+    cl = set()
+    with open(path, newline="") as fh:
+        rd = csv.reader(fh); next(rd)
+        for r in rd:
+            cl.add(r[4])
+    return cl
+cl_full = clusters_of(backup)
+cl_kept = clusters_of(list_csv)
+
+print(f"total chains in original list.csv : {rows_in}")
+print(f"chains removed                    : {removed_chains}")
+print(f"chains kept                       : {len(kept)}")
+print(f"unique PDB IDs removed            : {len(removed_pdbs)}")
+print(f"exclusion IDs present in dataset  : {len(matched_excl)} / {len(exclude)}")
+print(f"exclusion IDs NOT in dataset      : {len(exclude - matched_excl)}")
+print(f"clusters before / after           : {len(cl_full)} / {len(cl_kept)}  (fully dropped: {len(cl_full - cl_kept)})")

get_mhc_chains.py

@@ -0,0 +1,68 @@
+#!/usr/bin/env python
+"""Identify exact MHC chains in PDB via RCSB polymer-entity Pfam annotations,
+then map to the training clusters that must be dropped.
+
+MHC chain = a polymer entity annotated with an MHC-fold Pfam
+(PF00129 MHC-I/MHC-like incl CD1/MR1, PF00969 MHC-II-beta, PF00993 MHC-II-alpha),
+OR a chain whose sequence matches the (near-invariant) beta-2-microglobulin core
+(b2m carries a different Pfam, and b2m alone can anchor a pMHC assembly).
+"""
+import json, urllib.request, urllib.parse, sys, re, csv, os
+from collections import defaultdict
+
+MHC_PFAMS = {"PF00129", "PF00969", "PF00993"}
+B2M_SIG = re.compile(r"RHP[AP]ENGK..[FL]LNC")   # human/mouse b2m core, very conserved
+GQL = "https://search.rcsb.org/rcsbsearch/v2/query"
+DATA_GQL = "https://data.rcsb.org/graphql"
+
+def post_graphql(query):
+    req = urllib.request.Request(DATA_GQL, data=json.dumps({"query": query}).encode(),
+                                 headers={"Content-Type": "application/json"})
+    return json.load(urllib.request.urlopen(req, timeout=120))
+
+entries = [l.strip().lower() for l in open(sys.argv[1]) if l.strip()]   # mhc_pfam_pdbids.txt
+print(f"MHC-Pfam entries to resolve: {len(entries)}")
+
+mhc_chains = set()   # "pdbid_authchain", lowercase pdbid
+B = 80
+for i in range(0, len(entries), B):
+    batch = entries[i:i+B]
+    ids = ",".join(f'"{e.upper()}"' for e in batch)
+    q = f'''{{ entries(entry_ids:[{ids}]) {{
+        rcsb_id
+        polymer_entities {{
+          rcsb_polymer_entity_container_identifiers {{ auth_asym_ids }}
+          rcsb_polymer_entity_annotation {{ annotation_id type }}
+          entity_poly {{ pdbx_seq_one_letter_code_can }}
+        }} }} }}'''
+    try:
+        d = post_graphql(q)
+    except Exception as e:
+        print("batch error", i, e); continue
+    for ent in d["data"]["entries"]:
+        pdb = ent["rcsb_id"].lower()
+        for pe in ent.get("polymer_entities") or []:
+            anns = pe.get("rcsb_polymer_entity_annotation") or []
+            pfams = {a["annotation_id"] for a in anns if a.get("type") == "Pfam"}
+            seq = ((pe.get("entity_poly") or {}).get("pdbx_seq_one_letter_code_can") or "")
+            is_mhc = bool(pfams & MHC_PFAMS) or bool(B2M_SIG.search(seq))
+            if is_mhc:
+                for ch in (pe["rcsb_polymer_entity_container_identifiers"]["auth_asym_ids"] or []):
+                    mhc_chains.add(f"{pdb}_{ch}")
+    print(f"  resolved {min(i+B,len(entries))}/{len(entries)}  mhc_chains={len(mhc_chains)}")
+
+open("mhc_chains.txt","w").write("\n".join(sorted(mhc_chains))+"\n")
+print(f"total MHC chains identified: {len(mhc_chains)}")
+
+# map to training clusters
+DATA = sys.argv[2]
+rows = [r for r in csv.reader(open(os.path.join(DATA,"list.csv.full")))][1:]
+by_chain = {r[0].lower(): r for r in rows}
+mhc_clusters = set()
+hit = 0
+for ch in mhc_chains:
+    r = by_chain.get(ch)
+    if r: mhc_clusters.add(r[4]); hit += 1
+print(f"MHC chains present in training list: {hit}")
+print(f"MHC clusters to drop: {len(mhc_clusters)}")
+open("mhc_clusters.txt","w").write("\n".join(sorted(mhc_clusters))+"\n")

mhc_excluded_pdb_ids.txt

@@ -0,0 +1,1822 @@
+1a1m
+1a1n
+1a1o
+1a6a
+1a6z
+1a9b
+1a9e
+1agb
+1agc
+1agd
+1age
+1agf
+1akj
+1ao7
+1aqd
+1b0g
+1b0r
+1b3j
+1bd2
+1bii
+1bqh
+1bx2
+1bz9
+1c16
+1cd1
+1ce6
+1cg9
+1d5m
+1d5x
+1d5z
+1d6e
+1d9k
+1ddh
+1de4
+1dlh
+1duy
+1duz
+1e27
+1e28
+1ed3
+1eey
+1eez
+1efx
+1es0
+1exu
+1f3j
+1ffn
+1ffo
+1ffp
+1fg2
+1fne
+1fng
+1fo0
+1frt
+1fv1
+1fyt
+1fzj
+1fzk
+1fzm
+1fzo
+1g6r
+1g7p
+1g7q
+1gzp
+1gzq
+1h15
+1hdm
+1hhg
+1hhh
+1hhi
+1hhj
+1hhk
+1hla
+1hoc
+1hqr
+1hsa
+1hsb
+1hxy
+1hyr
+1i1a
+1i1f
+1i1y
+1i3r
+1i4f
+1i7r
+1i7t
+1i7u
+1iak
+1iao
+1iea
+1ieb
+1im3
+1im9
+1inq
+1j8h
+1je6
+1jf1
+1jgd
+1jge
+1jht
+1jk8
+1jl4
+1jpf
+1jpg
+1juf
+1jwm
+1jws
+1jwu
+1k2d
+1k5n
+1k8d
+1k8i
+1kbg
+1kcg
+1kg0
+1kj2
+1kj3
+1kjm
+1kjv
+1klg
+1klu
+1kpr
+1kpu
+1kpv
+1kt2
+1ktd
+1ktl
+1ld9
+1ldp
+1lds
+1leg
+1lek
+1lk2
+1lnu
+1lo5
+1lp9
+1m05
+1m6o
+1mhc
+1mhe
+1mi5
+1muj
+1mwa
+1n2r
+1n3n
+1n59
+1n5a
+1nam
+1nan
+1nez
+1of2
+1oga
+1ogt
+1onq
+1osz
+1p1z
+1p4l
+1p7q
+1pqz
+1py4
+1pyw
+1q94
+1qew
+1qlf
+1qo3
+1qqd
+1qr1
+1qrn
+1qse
+1qsf
+1qvo
+1r3h
+1r5i
+1r5v
+1r5w
+1rjy
+1rjz
+1rk0
+1rk1
+1s7q
+1s7r
+1s7s
+1s7t
+1s7u
+1s7v
+1s7w
+1s7x
+1s8d
+1s9v
+1s9w
+1s9x
+1s9y
+1seb
+1sje
+1sjh
+1sys
+1syv
+1t0m
+1t0n
+1t1w
+1t1x
+1t1y
+1t1z
+1t20
+1t21
+1t22
+1t5w
+1t5x
+1t7v
+1t7w
+1t7x
+1t7y
+1t7z
+1t80
+1tmc
+1tvb
+1tvh
+1u3h
+1u58
+1uqs
+1uvq
+1uxs
+1uxw
+1vac
+1vad
+1vgk
+1w0v
+1w0w
+1w72
+1wbx
+1wby
+1wbz
+1x7q
+1xh3
+1xr8
+1xr9
+1xz0
+1ydp
+1ymm
+1yn6
+1yn7
+1ypz
+1z5l
+1zag
+1zgl
+1zhb
+1zhk
+1zhl
+1zhn
+1zs8
+1zsd
+1zt1
+1zt4
+1zt7
+1zvs
+2a83
+2ak4
+2akr
+2av1
+2av7
+2axf
+2axg
+2bc4
+2bck
+2bnq
+2bnr
+2bsr
+2bss
+2bst
+2bvo
+2bvp
+2bvq
+2c7u
+2cii
+2cik
+2ckb
+2clr
+2clv
+2clz
+2d31
+2d4d
+2d4f
+2dyp
+2e7l
+2esv
+2f53
+2f54
+2f74
+2f8o
+2fik
+2fo4
+2fse
+2fwo
+2fyy
+2fz3
+2g9h
+2gaz
+2git
+2gj6
+2gt9
+2gtw
+2gtz
+2guo
+2h26
+2h6p
+2hjk
+2hjl
+2hla
+2hn7
+2iad
+2iam
+2ian
+2icw
+2ipk
+2j8u
+2jcc
+2mha
+2nna
+2nw3
+2nx5
+2oi9
+2oje
+2ol3
+2p24
+2p5e
+2p5w
+2po6
+2pxy
+2pye
+2q6w
+2q7y
+2qri
+2qrs
+2qrt
+2rfx
+2seb
+2uwe
+2v2w
+2v2x
+2vaa
+2vab
+2ve6
+2vlj
+2vlk
+2vll
+2vlr
+2wbj
+2wy3
+2x4n
+2x4o
+2x4p
+2x4q
+2x4r
+2x4s
+2x4t
+2x4u
+2x70
+2x89
+2xn9
+2xpg
+2ypk
+2ypl
+2yxf
+2z31
+2z9t
+2zok
+2zol
+2zsv
+2zsw
+3am8
+3arb
+3ard
+3are
+3arf
+3arg
+3au1
+3b3i
+3b6s
+3bgm
+3bh8
+3bh9
+3bhb
+3bo8
+3bp4
+3bp7
+3buy
+3bvn
+3bw9
+3bwa
+3bxn
+3bze
+3bzf
+3c5j
+3c5z
+3c60
+3c6l
+3c8k
+3c9n
+3cc5
+3cch
+3cdg
+3ch1
+3cii
+3cpl
+3cup
+3cvh
+3czf
+3d18
+3d25
+3d2u
+3d39
+3d3v
+3dbx
+3dhj
+3dhm
+3dmm
+3dtx
+3dx6
+3dx7
+3dx8
+3dxa
+3e2h
+3e3q
+3e6f
+3e6h
+3ecb
+3ekc
+3ery
+3es6
+3ffc
+3fol
+3fom
+3fon
+3fqn
+3fqr
+3fqt
+3fqu
+3fqw
+3fqx
+3fru
+3ft2
+3ft3
+3ft4
+3ftg
+3g08
+3giv
+3gjf
+3gml
+3gmm
+3gmn
+3gmo
+3gmp
+3gmq
+3gmr
+3gsn
+3gso
+3gsq
+3gsr
+3gsu
+3gsv
+3gsw
+3gsx
+3h7b
+3h9h
+3h9s
+3hae
+3hcv
+3he6
+3he7
+3hg1
+3hla
+3hpj
+3huj
+3i6g
+3i6k
+3i6l
+3ib4
+3ilp
+3ilq
+3ixa
+3jts
+3jtt
+3kla
+3kpl
+3kpm
+3kpn
+3kpo
+3kpp
+3kpq
+3kpr
+3kps
+3kww
+3kxf
+3kyn
+3kyo
+3l3d
+3l3g
+3l3h
+3l3i
+3l3j
+3l3k
+3l6f
+3lkn
+3lko
+3lkp
+3lkq
+3lkr
+3lks
+3ln4
+3ln5
+3low
+3lqz
+3lv3
+3m17
+3m1b
+3ma7
+3mbe
+3mgo
+3mgt
+3mr9
+3mrb
+3mrc
+3mrd
+3mre
+3mrf
+3mrg
+3mrh
+3mri
+3mrj
+3mrk
+3mrl
+3mrm
+3mrn
+3mro
+3mrp
+3mrq
+3mrr
+3mv7
+3mv8
+3mv9
+3myj
+3na4
+3nfn
+3nwm
+3o3a
+3o3b
+3o3d
+3o3e
+3o4l
+3o6f
+3o8x
+3o9w
+3ov6
+3ox8
+3oxr
+3oxs
+3p4m
+3p4n
+3p4o
+3p9l
+3p9m
+3pab
+3pdo
+3pgc
+3pgd
+3pl6
+3pqy
+3pwj
+3pwl
+3pwn
+3pwp
+3pwu
+3pwv
+3qda
+3qdg
+3qdj
+3qdm
+3qeq
+3qfd
+3qfj
+3qi9
+3qib
+3qiu
+3qiw
+3qq3
+3qq4
+3quk
+3qul
+3qux
+3quy
+3quz
+3qxa
+3qxd
+3qzw
+3rdt
+3rew
+3rgv
+3rl1
+3rl2
+3rol
+3roo
+3rtq
+3rug
+3rwc
+3rwd
+3rwe
+3rwf
+3rwg
+3rwh
+3rwi
+3rwj
+3rzc
+3s4s
+3s5l
+3s6c
+3scm
+3sda
+3sdc
+3sdd
+3sdx
+3sjv
+3skm
+3sko
+3spv
+3t0e
+3t1f
+3t8x
+3ta3
+3tbs
+3tbt
+3tbv
+3tbw
+3tby
+3tf7
+3tfk
+3tid
+3tie
+3tjh
+3tlr
+3tm6
+3tn0
+3to2
+3to4
+3tpu
+3tvm
+3tzv
+3u0p
+3ubx
+3upr
+3utq
+3uts
+3utt
+3v5d
+3v5h
+3v5k
+3vcl
+3vfm
+3vfn
+3vfo
+3vfp
+3vfr
+3vfs
+3vft
+3vfu
+3vfv
+3vfw
+3vh8
+3vj6
+3vri
+3vrj
+3vwj
+3vwk
+3vxm
+3vxn
+3vxo
+3vxp
+3vxr
+3vxs
+3vxu
+3w0w
+3w39
+3wl9
+3wlb
+3ws3
+3ws6
+3wuw
+3x11
+3x12
+3x13
+3x14
+4aen
+4ah2
+4apq
+4c56
+4d8p
+4e41
+4e5x
+4ei5
+4elm
+4en3
+4eup
+4f7m
+4f7p
+4f7t
+4fqx
+4ftv
+4fxl
+4g8g
+4g8i
+4g9d
+4g9f
+4gbx
+4gg6
+4gkn
+4gks
+4grl
+4gup
+4h1l
+4h25
+4h26
+4hkj
+4hs3
+4huu
+4huv
+4huw
+4hux
+4hv8
+4hwz
+4hx1
+4i0p
+4i48
+4i4w
+4i5b
+4iho
+4iiq
+4irj
+4irs
+4is6
+4jfd
+4jfe
+4jff
+4jfo
+4jfp
+4jfq
+4jqv
+4jqx
+4jrx
+4jry
+4k71
+4k7f
+4kdt
+4l29
+4l3c
+4l3e
+4l4t
+4l4v
+4l8b
+4l8c
+4l8d
+4l8s
+4l9l
+4lcc
+4lcw
+4lcy
+4lhu
+4lnr
+4m8v
+4may
+4mcy
+4mcz
+4md0
+4md4
+4md5
+4mdi
+4mdj
+4mj5
+4mj6
+4mji
+4mng
+4mnq
+4mq7
+4ms8
+4mvb
+4mx7
+4mxq
+4n0c
+4n0f
+4n0u
+4n5e
+4n8v
+4nhu
+4nnx
+4nny
+4no0
+4no2
+4no3
+4no5
+4nqc
+4nqd
+4nqe
+4nqv
+4nqx
+4nsk
+4nt6
+4o2c
+4o2e
+4o2f
+4ono
+4ov5
+4ozf
+4ozg
+4ozh
+4ozi
+4p23
+4p2o
+4p2q
+4p2r
+4p46
+4p4k
+4p4r
+4p57
+4p5k
+4p5m
+4p5t
+4pg2
+4pg9
+4pgb
+4pgc
+4pgd
+4pge
+4pj5
+4pj7
+4pj8
+4pj9
+4pja
+4pjb
+4pjc
+4pjd
+4pje
+4pjf
+4pjg
+4pjh
+4pji
+4pjx
+4pr5
+4pra
+4prb
+4prd
+4pre
+4prh
+4pri
+4prn
+4prp
+4pv8
+4pv9
+4qok
+4qrp
+4qrq
+4qrr
+4qrs
+4qrt
+4qru
+4r9h
+4ra3
+4rah
+4rmq
+4rmr
+4rms
+4rmt
+4rmu
+4rmv
+4rmw
+4s0u
+4u1h
+4u1i
+4u1j
+4u1k
+4u1l
+4u1m
+4u1n
+4u1s
+4u6x
+4u6y
+4uq2
+4uq3
+4wdi
+4wj5
+4wo4
+4wu5
+4wu7
+4wuu
+4ww2
+4wwk
+4x5w
+4x5x
+4x6c
+4x6d
+4x6e
+4x6f
+4xxc
+4y16
+4y19
+4y1a
+4y2d
+4y4f
+4y4h
+4y4k
+4z76
+4z77
+4z78
+4z7u
+4z7v
+4z7w
+4zak
+4zfz
+4zus
+4zut
+4zuu
+4zuv
+4zuw
+5b38
+5b39
+5bjt
+5brz
+5bs0
+5bxf
+5c07
+5c08
+5c09
+5c0a
+5c0b
+5c0c
+5c0d
+5c0e
+5c0f
+5c0g
+5c0i
+5c0j
+5c9j
+5cfh
+5cka
+5ckg
+5cs7
+5csb
+5csg
+5d2l
+5d2n
+5d5m
+5d7i
+5d7j
+5d7l
+5d9s
+5ddh
+5def
+5deg
+5dmk
+5e00
+5e6i
+5e8n
+5e8o
+5e8p
+5e9d
+5efi
+5enw
+5eo0
+5eo1
+5eot
+5eu3
+5eu4
+5eu5
+5eu6
+5euo
+5f1i
+5f1n
+5f7d
+5f9j
+5fa3
+5fa4
+5fdw
+5fkp
+5gr7
+5grd
+5grg
+5gsb
+5gsd
+5gsr
+5gsv
+5gsx
+5h94
+5hga
+5hgb
+5hgd
+5hgh
+5hhm
+5hhn
+5hho
+5hhp
+5hhq
+5hyj
+5ib1
+5ib2
+5ib3
+5ib4
+5ib5
+5ieh
+5iek
+5im7
+5inc
+5ind
+5iro
+5isz
+5iue
+5ivx
+5j1a
+5j6g
+5j6h
+5jhd
+5jlz
+5jwd
+5jwe
+5jzi
+5kd4
+5kd7
+5knm
+5ks9
+5ksa
+5ksb
+5ksu
+5ksv
+5l2j
+5l2k
+5lax
+5m00
+5m01
+5m02
+5men
+5meo
+5mep
+5meq
+5mer
+5mzm
+5n1y
+5n6b
+5nht
+5ni9
+5nig
+5nme
+5nmf
+5nmg
+5nmh
+5nmk
+5npz
+5nq0
+5nq1
+5nq2
+5nq3
+5nqk
+5opi
+5oqf
+5oqg
+5oqh
+5oqi
+5swq
+5sws
+5swz
+5t6w
+5t6x
+5t6y
+5t6z
+5t70
+5t7g
+5tez
+5til
+5tje
+5trz
+5ts1
+5tw2
+5tw5
+5txs
+5u16
+5u17
+5u1r
+5u2v
+5u6q
+5u72
+5u98
+5v4m
+5v4n
+5v5l
+5v5m
+5vcj
+5vcl
+5vgd
+5vge
+5vud
+5vue
+5vuf
+5vvp
+5vwd
+5vwf
+5vwh
+5vwj
+5vz5
+5w1v
+5w1w
+5w67
+5w69
+5w6a
+5wer
+5wes
+5wet
+5weu
+5whk
+5wjl
+5wjn
+5wke
+5wkf
+5wkg
+5wkh
+5wki
+5wl1
+5wlg
+5wli
+5wmn
+5wmo
+5wmp
+5wmq
+5wmr
+5wsh
+5wwi
+5wwj
+5wwu
+5wxc
+5wxd
+5xmf
+5xmm
+5xos
+5xot
+5xov
+5xs3
+5ylx
+5ymv
+5ymw
+5yxn
+5yxu
+6a6h
+6a97
+6aee
+6am5
+6amt
+6amu
+6apn
+6at5
+6at9
+6atf
+6ati
+6atz
+6avf
+6avg
+6bga
+6bij
+6bil
+6bin
+6bir
+6biv
+6bix
+6biy
+6biz
+6bj2
+6bj3
+6bj8
+6blq
+6blr
+6blx
+6bmh
+6bmk
+6bnk
+6bnl
+6bxp
+6bxq
+6c09
+6c15
+6c5m
+6c69
+6c6a
+6c6c
+6c6e
+6c6f
+6c6h
+6c6j
+6c97
+6c98
+6c99
+6cpl
+6cpn
+6cpo
+6cqj
+6cql
+6cqn
+6cqq
+6cqr
+6cug
+6cw6
+6cw9
+6cwb
+6cwe
+6cx5
+6cx7
+6cx9
+6cxa
+6cxe
+6cxf
+6cyw
+6d29
+6d2b
+6d2r
+6d2t
+6d64
+6d78
+6d7g
+6dfs
+6dfw
+6dfx
+6dig
+6dkp
+6e1i
+6ei2
+6eny
+6eqa
+6eqb
+6ewa
+6ewc
+6ewo
+6fgb
+6g3j
+6g3k
+6g9q
+6g9r
+6gb5
+6gb6
+6gb7
+6ggm
+6gh1
+6gh4
+6ghn
+6gl1
+6h6d
+6h6h
+6hby
+6i8c
+6id4
+6iex
+6ilc
+6ile
+6ilf
+6ilg
+6ilm
+6iwg
+6iwh
+6j1v
+6j1w
+6j29
+6j2a
+6j2d
+6j2e
+6j2f
+6j2g
+6j2h
+6j2i
+6j2j
+6joz
+6jp3
+6jq2
+6jq3
+6jtn
+6jto
+6jtp
+6k60
+6k7t
+6k7u
+6kwk
+6kwl
+6kwn
+6kwo
+6l9k
+6l9l
+6l9m
+6l9n
+6la6
+6la7
+6lah
+6lam
+6lb2
+6lf8
+6lf9
+6lt6
+6m1b
+6m24
+6m2j
+6m2k
+6mff
+6mfg
+6miv
+6miy
+6mj4
+6mj6
+6mja
+6mji
+6mjj
+6mjq
+6mkd
+6mkr
+6mng
+6mnm
+6mnn
+6mno
+6mp0
+6mp1
+6mpp
+6mss
+6mt3
+6mt4
+6mt5
+6mt6
+6mtl
+6mtm
+6mwr
+6nca
+6nf7
+6nha
+6nix
+6npr
+6nux
+6o4y
+6o4z
+6o51
+6o53
+6o9b
+6o9c
+6ojp
+6omg
+6oor
+6opd
+6p23
+6p27
+6p2c
+6p2f
+6p2s
+6p64
+6pa1
+6pag
+6pbh
+6ptb
+6pte
+6puc
+6pud
+6pue
+6puf
+6pug
+6puh
+6pui
+6puj
+6puk
+6pul
+6pum
+6pvc
+6pvd
+6px6
+6py2
+6pyj
+6pyl
+6pyv
+6pyw
+6pz5
+6q3k
+6q3s
+6qio
+6qip
+6qza
+6qzc
+6qzd
+6r0e
+6r2l
+6r2u
+6rp9
+6rpa
+6rpb
+6rsy
+6ss7
+6ss8
+6ss9
+6ssa
+6tdo
+6tdp
+6tdq
+6tdr
+6tds
+6tmo
+6trn
+6tro
+6u3m
+6u3n
+6u3o
+6uj7
+6uj8
+6uj9
+6ujo
+6ujq
+6uk2
+6uk4
+6uli
+6ulk
+6uln
+6ulr
+6uon
+6uz1
+6uzm
+6uzn
+6uzo
+6uzp
+6uzq
+6uzs
+6v0y
+6v13
+6v15
+6v18
+6v19
+6v1a
+6v2o
+6v2p
+6v2q
+6v3j
+6v7y
+6v7z
+6v80
+6vb0
+6vb1
+6vb2
+6vb3
+6vb4
+6vb5
+6vb6
+6vb7
+6viu
+6vm7
+6vm8
+6vm9
+6vma
+6vmc
+6vmx
+6vpz
+6vq2
+6vqd
+6vqe
+6vqo
+6vqy
+6vqz
+6vr1
+6vr5
+6vrm
+6vrn
+6w51
+6w9u
+6w9v
+6wl2
+6wl3
+6wl4
+6wna
+6wzy
+6x00
+6xc9
+6xco
+6xcp
+6xng
+6xp6
+6xqa
+6xqp
+6y26
+6y27
+6y28
+6y29
+6y2a
+6y2b
+6z9v
+6z9w
+6z9x
+6zkw
+6zkx
+6zky
+6zkz
+6zwa
+7afv
+7alo
+7bbg
+7bh8
+7byd
+7c9v
+7cjq
+7cpo
+7dc6
+7dzm
+7dzn
+7fi5
+7fi6
+7fi7
+7fi8
+7fi9
+7ji2
+7jwi
+7jwj
+7jyu
+7jyv
+7jyw
+7jyx
+7k80
+7k81
+7kei
+7kgo
+7kgp
+7kgq
+7kgr
+7kgs
+7kgt
+7koz
+7kp0
+7kp1
+7l1b
+7l1c
+7l1d
+7lfi
+7lfj
+7lfk
+7lfl
+7lfm
+7lfz
+7lg0
+7lg2
+7lg3
+7lgd
+7lgt
+7lli
+7llj
+7m72
+7m8s
+7m8t
+7m8u
+7mj6
+7mj7
+7mj8
+7mj9
+7mkb
+7mle
+7n19
+7n1a
+7n1b
+7n1e
+7n1f
+7n2n
+7n2o
+7n2p
+7n2q
+7n2r
+7n2s
+7n4k
+7n5c
+7n5p
+7n5q
+7n6d
+7n6e
+7na5
+7ndq
+7ndt
+7ndu
+7nme
+7nmf
+7nmg
+7nui
+7nze
+7o00
+7ow5
+7ow6
+7p0a
+7p0t
+7p3d
+7p3e
+7p49
+7p4b
+7pb2
+7pbc
+7pbe
+7pdw
+7phr
+7q15
+7q98
+7q99
+7q9a
+7q9b
+7qhp
+7qng
+7qpd
+7qpj
+7r80
+7rdv
+7re7
+7re8
+7rk7
+7rm4
+7rrg
+7rtd
+7rtr
+7rzd
+7rzj
+7s79
+7s7d
+7s7e
+7s7f
+7s8a
+7s8e
+7s8f
+7sa2
+7sg0
+7sg1
+7sg2
+7t2a
+7t2b
+7t2c
+7t2d
+7tlt
+7tr4
+7tuc
+7tud
+7tue
+7u1r
+7u21
+7uc5
+7ufj
+7um2
+7umg
+7ur1
+7wcy
+7xf3
+7xxa
+7yx9
+7yxb
+7z0q
+7z50
+7zt2
+7zt3
+7zt4
+7zt5
+7zt7
+7zt8
+7zt9
+8bp6
+8cmb
+8cmc
+8cmd
+8cme
+8cmf
+8cmg
+8cmh
+8cmi
+8cx4
+8d5e
+8d5f
+8d5j
+8d5k
+8d5n
+8d5q
+8dnt
+8dv4
+8dvg
+8e13
+8e2z
+8e8i
+8eb2
+8en8
+8enh
+8eo8
+8es8
+8es9
+8esa
+8esb
+8euq
+8f5a
+8fhl
+8fhu
+8fu4
+8gom
+8gon
+8gvb
+8gvg
+8gvi
+8hsm
+8hso
+8hsw
+8ht1
+8ht9
+8i5c
+8i5d
+8jhw
+8k4t
+8k4v
+8k50
+8pje
+8pjf
+8pjg
+8qfy
+8ral
+8rbv
+8rlt
+8rlu
+8rlv
+8rng
+8rnh
+8rni
+8roo
+8rop
+8rro
+8rwb
+8rym
+8ryo
+8ryp
+8sgb
+8shi
+8t4z
+8tbp
+8tbv
+8tbw
+8tlz
+8tm0
+8tm2
+8tq4
+8tq7
+8tq8
+8tq9
+8tqa
+8trl
+8trq
+8trr
+8tub
+8tuh
+8u9g
+8udr
+8umo
+8utc
+8v8q
+8vcl
+8vcx
+8vcy
+8vd0
+8vd2
+8vjz
+8vq8
+8vr9
+8vra
+8vrw
+8vsj
+8vsp
+8vz8
+8vz9
+8w83
+8w84
+8w85
+8w86
+8wte
+8wul
+8xes
+8xfz
+8xg2
+8xkc
+8y6x
+8ye4
+8z05
+9asf
+9asg
+9aud
+9b7b
+9bf9
+9btx
+9bty
+9btz
+9bu0
+9bys
+9c42
+9c9d
+9cgr
+9cgs
+9cx0
+9cx1
+9cx2
+9cyl
+9cym
+9d72
+9dy8
+9ejg
+9ejh
+9eji
+9ek6
+9ek7
+9hi7
+9hy4
+9j4v
+9l1l
+9min
+9ms0
+9nav
+9nfc
+9nig
+9nih
+9o05
+9o06
+9o07
+9o08
+9uv8
+9wbd
+9wve
+9wvf
+9y0t
+9y0u
+9y0v
+9y0w

proteinmpnn_nomhc.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e46fe0a014231183c4571dc368daa0b884b78c622b5986b3416de876f8230e75
+size 20067862