Poster Presentation 47th Lorne Genome Conference 2026

Comprehensive Evaluation of ACMG/AMP-based Variant Classification Tools (131950)

Tohid Ghasemnejad 1 2 , Yuheng Liang 1 , Khadijeh Hoda Jahanian 3 , Milad Eidi 4 , Arash Salmaninejad 5 , Seyedeh Sedigheh Abedini 1 , Fabrizzio Horta 6 , Nigel H Lovell 1 , Thantrira Porntaveetus 3 7 , Mark Grosser 7 , Mahmoud Aarabi 8 , Hamid Alinejad-Rokny 1
  1. UNSW BioMedical Machine Learning Lab (BML), School of Biomedical Engineering, UNSW Sydney, Sydney, NSW 2052, Australia, Kensington, NSW, Australia
  2. 1UNSW BioMedical Machine Learning Lab (BML), School of Biomedical Engineering, UNSW Sydney, Sydney, NSW 2052, Australia, NSW
  3. Center of Excellence in Precision Medicine and Digital Health, Department of Physiology, Faculty of Dentistry, Chulalongkorn University, Bangkok, Thailand, Bangkok, Thailand
  4. The International ImMunoGeneTics Information System (IMGT), National Center for Scientific Research (CNRS), Institute of Human Genetics (IGH), University of Montpellier (UM), Montpellier, France, France
  5. Center for Individualized Medicine, Mayo Clinic, Rochester, MN 55905, USA, Rochester
  6. Fertility & Research Centre, Discipline of Women's Health, School of Clinical Medicine & The Royal Hospital for Women, UNSW Sydney, Randwick, Australia, Kensington, NSW, Australia
  7. 23Strands, Pyrmont, Australia, Sydney, NSW, Australia
  8. Departments of Pathology and Obstetrics, Gynecology and Reproductive Sciences, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, USA, Pennsylvania, USA

Background: The American College of Medical Genetics and Genomics/Association for Molecular Pathology (ACMG/AMP) guidelines represent the gold standard for clinical variant interpretation. Despite widespread adoption, a comprehensive comparison of software tools designed to implement these guidelines has been lacking, creating a significant gap in evidence-based guidance for clinicians.

Methods: We benchmarked four ACMG/AMP-based tools (Franklin, InterVar, TAPES, Genebe) selected from 22 available tools, and compared their performance with LIRICAL, a top-performing phenotype-driven tool. Selection criteria included free availability, VCF compatibility, operational reliability, and absence of disease-specificity. We used 151 expert-curated datasets from Mendelian disorders. Our evaluation framework assessed top-N accuracy (N=1,5,10,20,50), retention rates, precision, recall, F1 scores, and Area Under the Curve (AUC). Statistical validation employed bootstrap confidence intervals (n=1000) and Friedman tests.

Results: LIRICAL (68.21%) and Franklin (61.59%) demonstrated superior top-10 variant prioritization accuracy in Mendelian disorders, significantly outperforming other tools (p=0.0000).

Conclusions: Tools with advanced phenotypic integration significantly outperform those relying primarily on genomic features for variant prioritization in Mendelian disorders.