Curating the Clinical Genome 2024 ePoster Gallery-ClinVar-Based Automated Classifier Accelerates Variant Interpretation for Cancer Predisposition Syndromes

ClinVar-Based Automated Classifier Accelerates Variant Interpretation for Cancer Predisposition Syndromes

Pdf Summary

The Mayo Foundation for Medical Education and Research developed a two-step approach to construct a binary variant classifier for cancer predisposition syndromes. The first step involved computing concordance rates with ClinGen curations for ClinVar submitters. The second step included constructing the classifier using data from ClinVar submitters with ClinGen concordance rates above a given threshold. The objective was to evaluate the performance of the classifier using a validation dataset, where it showed no misclassifications and accurately classified P/LP and non-P/LP variants. The study found that the ClinVar-based automated variant classifiers significantly reduced the manual interpretation burden while maintaining accuracy, eliminating manual interpretation for 95.8% of the variants in the Mayo validation dataset. The classifiers are particularly useful for scenarios where only P/LP variants are reported, such as in secondary findings. Further studies are needed to determine the generalizability of these results across different hereditary conditions. Overall, the automatic classification variants showed high positive predictive values (PPV) and negative predictive values (NPV) in accurately classifying variants.

Keywords

Mayo Foundation

Medical Education

Research

Variant Classifier

Cancer Predisposition Syndromes

ClinGen

ClinVar

Concordance Rates

Validation Dataset

Automated Variant Classification