EUVIMEDOpdag EUVIMED
Open access · CC BY 4.0

The impact of curation errors in the PDBBind Database on machine learning predictions of protein–protein binding affinity

Jason D Gans iD, Justin E Miller, Esen Sokullu, Nileena Velappan iD, Ramesh K Jha iD

Database · 2025

DOI: 10.1093/database/baaf061 ↗

At a glance

Journal article in Database, 2025. Written by Jason D Gans and 4 co-authors. Institution: Bioscience Division, Los Alamos National Laboratory, Los Alamos, NM, 87545, USA. Major MEDLINE topics: Machine Learning, Databases, Protein, Proteins, Data Curation. Cited 3 times so far (Crossref); 45 references. The full text is freely available (open access, CC BY 4.0). Funded by: Laboratory Directed Research and Development.

AbstractSubjectsBibliografiske oplysningerCiteRelated studies

Abstract (original language)

The PDBBind database has been widely utilized for the computational prediction of protein–protein binding affinities. While the accuracy of the PDBBind-curated equilibrium dissociation constants (KD) has been reported for the protein–ligand subset of the PDBBind database, the curation accuracy has not been reported for the protein–protein subset. Here, we present a detailed manual analysis for the subset of PDBBind records with PubMed Central Open Access primary publications and find that ~19% of these records had KD values that were not supported by their primary publications. The impact of these putative curation errors on the machine learning-based prediction of KD from experimental protein–protein 3D structures was evaluated and correcting the curation errors improved the Pearson correlation coefficient between measured and random forest-predicted log10(KD) values by ~8 percentage points. This finding underscores the importance of dataset accuracy for computational modelling and highlights the need for more stringent curation processes when extracting information from the scientific literature.

Subjects

MeSH terms (MEDLINE)

Machine LearningDatabases, ProteinProteinsData CurationProtein Binding

Substances

Proteins

Authors and institutions

  • Bioscience Division, Los Alamos National Laboratory, Los Alamos, NM, 87545, USA.

ORCID profiles: Jason D Gans, Nileena Velappan, Ramesh K Jha

Funding

  • Laboratory Directed Research and Development

Cite

APA

Gans, J. D., Miller, J. E., Sokullu, E., Velappan, N., & Jha, R. K. (2025). The impact of curation errors in the PDBBind Database on machine learning predictions of protein–protein binding affinity. Database. https://doi.org/10.1093/database/baaf061

Vancouver

Gans JD, Miller JE, Sokullu E, Velappan N, Jha RK. The impact of curation errors in the PDBBind Database on machine learning predictions of protein–protein binding affinity. Database. 2025. doi:10.1093/database/baaf061

BibTeX
@article{Gans2025,
  author = {Gans, Jason D and Miller, Justin E and Sokullu, Esen and Velappan, Nileena and Jha, Ramesh K},
  title = {The impact of curation errors in the PDBBind Database on machine learning predictions of protein–protein binding affinity},
  journal = {Database},
  year = {2025},
  doi = {10.1093/database/baaf061}
}
Start for free →