IF-Split-2026.07.22: fold-aware PDB train/validation/test split and pinned metadata snapshot

The pinned PDB metadata snapshot and train/validation/test split behind the preprint"Fold Leakage in Inverse-Folding Benchmarks, and the Limit on Fixing It: Auditing theProteinMPNN and LigandMPNN Splits". Everything here is metadata only: sequences, sequence-cluster memberships, fold anddomain-family assignments, and ligand signals. No structure coordinates. That is thepoint of the method, since the split is computed without downloading a single mmCIFfile, which makes it megabytes of JSON rather than a terabyte of structures. Code: https://github.com/WSobo/IF-Split (MIT license, tag v0.6.1) CONTENTS candidates-annotated.jsonl (471 MB) - the pinned snapshot. 225,653 PDB entries, oneJSON record each, including Pfam and InterPro domain families. dataset.lock (3.2 MB) - pins the candidate set by sha256, the full configuration, andthe hash of the split output. manifest.json (5.4 KB) - the reported statistics for this split. train.json, val.json, test.json (1.7 MB total) - the split itself, as lists of PDBentry ids. metal_test.json, nucleic_acid_test.json, small_molecule_test.json - theligand-class-stratified test sets. cfg-recommended.yaml (890 B) - the exact configuration that produced this split. Itsconfig_hash is f7d4203586df3dc7b10d2948e76d20d8, which dataset.lock also records. SHA256SUMS - checksums for everything above. It was written against the layout thesefiles have in the build tree, where the test lists sit under "test/" and the config iscalled config.yaml. The upload flattened the first and kept the config's original name,so rewrite those two before checking: sed -e 's| test/| |' -e 's| config.yaml| cfg-recommended.yaml|' SHA256SUMS | sha256sum -c All ten lines should report OK. REPRODUCING THE SPLIT pip install if-split==0.6.1 if-split resplit --candidates candidates-annotated.jsonl --config cfg-recommended.yaml --out ./out if-split verify dataset.lock --candidates candidates-annotated.jsonl The last command re-derives the split offline, with no network access, and checks itagainst the lock. It should print: OK: reproduced exactly (225653 entries; candidates + split verified). It exits non-zero and prints "DRIFT detected" if either the snapshot or the derivedsplit differs, so it is safe to use as a continuous-integration gate. THE SPLIT Merging on all five authorities (CATH, ECOD, SCOP2, Pfam, InterPro) with thedata-derived "maximal" strategy: train: 213,890 entries, in 1 leakage-safe component validation: 1,467 entries, in 813 components test: 1,465 entries, in 864 components Training is a single component. That is not a configuration choice, it is the finding.Merging entities that share a family collapses the PDB into one component holding 98.6%of entries, so at most 1.35% can be held out at the fold level, and this split takesessentially all of it. Of the 2,932 held-out entries, 888 are certified novel: they carry a family under atleast one of the five authorities and share none of them with training. The remaining2,044 carry no annotation under any authority. They are fold-disjoint but unverifiable,and are reported separately rather than counted as novel. CAVEATS WORTH READING FIRST Only the annotated snapshot is included, deliberately. An unannotated candidates filecarries no Pfam or InterPro families, cannot reproduce any of the result above, andfails verification against this lock. The analysis is snapshot-dependent in a known direction. CATH, ECOD and SCOP2 arecurated retrospectively from solved structures, so their coverage grows over time. Asthey annotate the recent PDB, more chains acquire merge edges and the fold-disjointresidual shrinks. The 1.35% here is an upper bound on what a later snapshot will yield. Domain-family disjointness is not fold disjointness. Sharing an InterPro family is aclaim about a domain rather than about a fold. The holdout is disjoint under thecriteria named above; residual leakage below those criteria is bounded and disclosed,not zero. CITATION Cite the preprint for the analysis and this DOI for the snapshot. The lock embeds thefull configuration, so anyone holding only dataset.lock can recover the exact settingswithout this description.