An annotation-overlap-flagged rare-disease gene-prioritisation benchmark and PMC index recipe
- 1Universidad Europea de Valencia
Abstract
Benchmarks for rare-disease gene prioritisation are assembled from published clinical cases. Those cases often come from the same publications used to build knowledge-base ("curated") tools, so a curated tool can be scored on its own source literature. This resource makes that circularity measurable. We release a stratified benchmark of 1,047 rare-disease cases from the GA4GH Phenopacket Store v0.1.26. Each case pairs a Human Phenotype Ontology profile with a 50-gene candidate list (one causal gene, 49 distractors) and the causal-gene label, sampled across four operational MONDO-derived disease strata and issued in two case-paired variants: random distractors, and phenotype-similar distractors selected by HPO Resnik similarity. Two case-level metadata layers support fairer evaluation: a per-case flag recording whether a case's source publication is cited in the HPO disease-annotation file, defining an overlap-absent subset (n = 282), and publication-recency strata. We also specify a deterministic, version-pinned recipe for a hybrid dense-plus-sparse retrieval index over approximately 2.25 million PMC Open Access articles (52,777,395 chunks). The resource reports no tool comparisons.
Related
- Where single-cell foundation models help: a benchmark of annotation, perturbation and regulatory inference against simple baselines
- Cell-type separability predicts annotation accuracy and outweighs algorithm choice: a factorial benchmark across seven scRNA paradigms
- MillionTrees: A global benchmark for tree detection across geography, sensor, and annotation geometry
- Understanding and translating evolutionary adaptations in animals exposed to contaminated diets to guide strategies against human foodborne disease: a prioritisation framework
- Preprocessing strategies for plant disease instance segmentation: annotation efficiency, context signals, and dual-metric evaluation on PlantSeg.
BibTeX
@misc{angulo2026annotationoverlapflagged,
title = {{An annotation-overlap-flagged rare-disease gene-prioritisation benchmark and PMC index recipe}},
author = {J. Angulo and V. Yeste and H. Espinos-Morato},
howpublished = {bioRxiv},
year = {2026},
month = aug,
doi = {10.64898/2026.08.30.748028},
url = {https://www.biorxiv.org/content/10.64898/2026.08.30.748028v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - An annotation-overlap-flagged rare-disease gene-prioritisation benchmark and PMC index recipe
AU - J. Angulo
AU - V. Yeste
AU - H. Espinos-Morato
PY - 2026
DA - 2026/08/30
PB - bioRxiv
DO - 10.64898/2026.08.30.748028
UR - https://www.biorxiv.org/content/10.64898/2026.08.30.748028v1
AB - Benchmarks for rare-disease gene prioritisation are assembled from published clinical cases. Those cases often come from the same publications used to build knowledge-base ("curated") tools, so a curated tool can be scored on its own source literature. This resource makes that circularity measurable. We release a stratified benchmark of 1,047 rare-disease cases from the GA4GH Phenopacket Store v0.1.26. Each case pairs a Human Phenotype Ontology profile with a 50-gene candidate list (one causal gene, 49 distractors) and the causal-gene label, sampled across four operational MONDO-derived disease strata and issued in two case-paired variants: random distractors, and phenotype-similar distractors selected by HPO Resnik similarity. Two case-level metadata layers support fairer evaluation: a per-case flag recording whether a case's source publication is cited in the HPO disease-annotation file, defining an overlap-absent subset (n = 282), and publication-recency strata. We also specify a deterministic, version-pinned recipe for a hybrid dense-plus-sparse retrieval index over approximately 2.25 million PMC Open Access articles (52,777,395 chunks). The resource reports no tool comparisons.
ER -