遇见数据集

TMRDis reference dataset: residue-level disorder labels for α-helical transmembrane proteins

收藏
Zenodo2026-11-04 更新2026-11-05 收录
官方服务:

资源简介:

Training and evaluation reference of TMRDis, a predictor of intrinsically disordered residues in α-helical transmembrane proteins. The dataset covers 5,067 reviewed UniProtKB/Swiss-Prot α-helical transmembrane proteins (4,997 distinct sequences, 3,303,306 residues). They are divided into training, validation and test partitions (3,547 / 760 / 760 proteins) by 30%-identity sequence components. Each residue has a merged disorder label and the evidence that set it: expert-curated annotations (DisProt, IDEAL and UniProt, through MobiDB), NMR chemical-shift scores (TriZOD) or missing and observed residues in PDB structures (MobiDB). Positions without eligible evidence are marked unknown: 1,796,978 residues have a label (284,621 disorder, 1,512,357 non-disorder) and 1,506,328 are unknown. The record also contains residue labels for the MemDis and TriZOD test sets, and the versions and thresholds of the 12 published predictors compared with TMRDis plus the PUNCH2-Light configuration. README.md describes the files and columns; ATTRIBUTION.md lists the sources and their licences. TMRDis code: https://github.com/sunnylab-cb/tmrdis

提供机构:
Zenodo
创建时间:
2026-11-04
二维码
社区交流群
二维码
科研交流群
商业服务