Comprehensive evaluation of ACMG/AMP-based variant classification tools

N/A (151 Mendelian disease VCFs)

First comprehensive benchmark of 4 ACMG/AMP classification tools (Franklin, InterVar, TAPES, Genebe) on 151 real clinical VCFs from Mendelian disease patients, curated by two independent clinical geneticists; evaluation of causal variant prioritization

First comprehensive benchmark of four ACMG/AMP classification tools (Franklin, InterVar, TAPES, Genebe) evaluated on 151 real clinical VCFs from Mendelian disease patients, curated by two independent clinical geneticists. Franklin and LIRICAL consistently outperform others in causal variant prioritization. InterVar, TAPES and Genebe show inferior performance in this real clinical context, particularly for false-positive variants. Performance varies by disease and gene.

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

Benchmark on real Mendelian disease VCFs is far more representative than ClinVar-only evaluation. Franklin leading is consistent with its AI integration and regular updates. Multi-tool evaluation reveals that performance varies by disease: a multi-tool pipeline (Franklin + LIRICAL) is recommended in clinical routine. Reassess InterVar if still used as a single tool.

benchmark on real VCFs (vs ClinVar alone) +2; direct impact on tool selection in routine +2; Franklin leading (actionable result) +2; Bioinformatics (Oxford) +1