ECOD v295.2: Evolutionary Classification of Protein Domains, release data

A fixed, citable snapshot of the core distributable files for release v295.2 of ECOD, the Evolutionary Classification of Protein Domains. This release classifies 2,943,252 domains from experimental structures in the Protein Data Bank and from predicted structures, into 2,456 X-groups, 3,715 H-groups, 3,951 T-groups and 32,042 F-groups, with the family level defined in collaboration with Pfam 38.2. The deposit contains the domain table, the classification hierarchy, the F-group to Pfam accession mapping, ECOD's published md5 manifest and the release notes. It deliberately does not mirror the full distribution: coordinate archives, sequence and structure search databases, and the moving ecod.latest.* series are not included, and remain available from the live resource. ECOD is maintained at http://prodata.swmed.edu/ecod, where every release since 2015 is archived. This deposit exists to fix the exact contents of v295.2 as described in the accompanying publication, and to allow the counts reported there to be recomputed independently. Every file was verified against ECOD's own published md5 manifest at the time of deposit; both the original and the compressed checksums are recorded in MANIFEST.json. Correction (version 295.2-r1). In the original deposit the ligand-contact columns of ecod.v295.2.domains.txt (ligand_binding, ligand_comp_ids, ligand_pdbnum) were empty for every domain because of a defect in the release process. This version repopulates them: 545,894 domains carry ligand contacts (within 4.0 Å). No domain, range, classification or other column has changed. ecod.v295.2.md5 and MANIFEST.json are updated to match; details are in CHANGES.md.