A topology census of four-strand β-sheets in experimental and predicted protein structures: code and data
Year: 2026
Type: dataset
Abstract: Code and data for A topology census of four-strand β-sheets in experimental and predicted protein structures. Ruczinski et al. (2002) enumerated the 96 topologies of the four-strand β-sheet and reported that 48 were unobserved in the contemporary PDB. This deposition contains the artifacts of a re-measurement of that result with a modernized implementation of the ProSMoS pattern search, applied uniformly to 1.4 million searched experimental ECOD domains (of an archive of 1,491,961) and to 4.91 million DPAM-parsed representative domains from a provisional classification of the AlphaFold database 200M set. It includes: the full 96-topology enumeration with its canonical encoding and the hand transcription of the 2002 figure it is checked against; per-topology occupancy in both archives at three containment conventions (the topology is the whole sheet; it is embedded in a sheet of at most six strands; it is embedded in a sheet of any size); the rarefaction, monotypy and confinement analyses over ECOD X-groups; every control and sensitivity analysis reported in the paper's Supporting Information, including an orthogonal DSSP hydrogen-bond detector, a geometric acceptance-threshold sweep and a cardinality sweep; the frozen ECOD v295 classification snapshot that every X-group number is computed against; all analysis scripts; and the rendered figures with their generators. A same-molecule idealization control is also deposited (global_e_geom.*, pairs.json, agreement_sse_corrected.json); it was computed on the superseded acceptance rule and its Supporting Information section was withdrawn before submission, so it is shipped as evidence rather than as a reported result. Neither the manuscript nor its Supporting Information is deposited here; this record is the analysis behind the paper, not the paper. Section numbers in the README refer to the Supporting Information published with the article. A complete manifest with per-file descriptions, SHA-256 checksums, and the provenance of every number is in README.md, which is the entry point for this record. ADOPTION_PLANE.md states the geometric acceptance rule the census uses and what changed when it was adopted, and should be read before trusting any individual artifact. The ProSMoS search engine is not in this deposition; it is archived separately at 10.5281/zenodo.22075371 (v1.0.0), and the PALSSE secondary-structure parser at 10.5281/zenodo.22074592 (v2.1.0). The predicted domain structures the census was run over are a companion record, 10.5281/zenodo.22074674. This deposition is the analysis: everything downstream of the search.
Institution: The University of Texas Southwestern Medical Center
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7204107869