EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

BioGraphBench-PPI enables audit-first evaluation of negative-sampling contracts and structural baselines in protein interaction networks

Alejandro Cabrera-Andrade, Jorge Robalino-Díaz, Iván Dueñas-Espín, Ana Karina Zambrano, William Villegas-Ch

Frontiers in Bioinformatics · 01.10.2026 · 6

DOI: 10.3389/fbinf.2026.1966469 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Frontiers in Bioinformatics, online erschienen am 01.10.2026. Verfasst von Alejandro Cabrera-Andrade und 4 weiteren Autor:innen. Beteiligte Einrichtung: Grupo de Bio-Quimioinformática, Universidad de Las Américas, Quito, Ecuador. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 24 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

Biomedical graph-learning results are difficult to compare when source versions, identifier rules, data splits, structural features, and sampled-nonedge contracts are incompletely documented.

Methods

We developed BioGraphBench-PPI as an audit-first benchmark for protein-protein interaction (PPI) link prediction using STRING v12.0, BioGRID 5.0.260, and two cross-resource no-overlap ablations. For each PPI task, we generated 10 predefined seeds and three balanced sampled-nonedge contracts: random, degree-matched, and two-hop. We compared local heuristics, logistic regression, random forest, histogram gradient boosting (HGB), an auditable node2vec-compatible random-walk baseline, and a two-layer GCN link-prediction baseline, reporting ranking, calibration, runtime, and confidence intervals.

Results

Random sampled nonedges yielded high HGB AUPRC values (0.942–0.964), but degree-matched sampled nonedges reduced AUPRC by 0.049–0.259. The GCN baseline showed signal under random negatives in selected graphs (AUPRC 0.526–0.717) but approached chance under degree-matched and two-hop contracts (AUPRC 0.500–0.515).

Discussion

BioGraphBench-PPI shows that PPI benchmark conclusions depend strongly on sampled-nonedge design and graph source. The benchmark therefore contributes an auditable evaluation contract rather than a claim that PPI prediction is solved.

Schlagwörter

Stichwörter der Autor:innen

benchmarkingbiomedical networkscalibrationlink predictionnegative samplingnode2vecprotein-protein interactionsreproducibility

Autor:innen und Institutionen

  • Grupo de Bio-Quimioinformática, Universidad de Las Américas, Quito, Ecuador.
  • Escuela de Enfermería, Facultad de Ciencias de La Salud, Universidad de Las Américas, Quito, Ecuador.
  • Escuela de Ingeniería en Ciberseguridad, Facultad de Ingeniería y Ciencias Aplicadas, Universidad de Las Américas, Quito, Ecuador.
  • Grupo de Investigación en Inequidades y Determinantes Ecológicos y Ambientales en Salud, Instituto de Salud Pública, Facultad de Salud y Bienestar, Pontificia Universidad Católica del Ecuador, Quito, Ecuador.
  • Facultad de Ciencias de La Salud Eugenio Espejo, Centro de Investigación Genética y Genómica, Universidad UTE, Quito, Ecuador.

Zitieren

APA

Cabrera-Andrade, A., Robalino-Díaz, J., Dueñas-Espín, I., Zambrano, A. K., & Villegas-Ch, W. (2026). BioGraphBench-PPI enables audit-first evaluation of negative-sampling contracts and structural baselines in protein interaction networks. Frontiers in Bioinformatics, 6. https://doi.org/10.3389/fbinf.2026.1966469

Vancouver

Cabrera-Andrade A, Robalino-Díaz J, Dueñas-Espín I, Zambrano AK, Villegas-Ch W. BioGraphBench-PPI enables audit-first evaluation of negative-sampling contracts and structural baselines in protein interaction networks. Frontiers in Bioinformatics. 2026;6. doi:10.3389/fbinf.2026.1966469

BibTeX
@article{CabreraAndrade2026,
  author = {Cabrera-Andrade, Alejandro and Robalino-Díaz, Jorge and Dueñas-Espín, Iván and Zambrano, Ana Karina and Villegas-Ch, William},
  title = {BioGraphBench-PPI enables audit-first evaluation of negative-sampling contracts and structural baselines in protein interaction networks},
  journal = {Frontiers in Bioinformatics},
  year = {2026},
  volume = {6},
  doi = {10.3389/fbinf.2026.1966469}
}
Kostenlos starten →