Quantitative modeling: multifactorial integration of data

Abstract Quantitative approaches can be used for the clinical interpretation of variants of uncertain significance. Evidence used in variant interpretation as negative or positive predictors can be converted into quantitative statistical measures. Multifactorial likelihood analysis is a statistical method for integrating different data types using Bayesian modeling. A prior probability of pathogenicity is updated with combined odds or likelihood ratios for pathogenicity to derive a posterior probability of pathogenicity that is used to determine the classification of a given variant. Using a well-characterized reference set of (assumed) pathogenic and (assumed) benign, various techniques can be used to derive components of the model from categorical and continuous data. This type of modeling has been applied to the American College of Medical Genetics and Genomics and the Association of Molecular Pathology (ACMG/AMP) variant classification guidelines, and various hereditary cancer genes, including BRCA1/2, mismatch repair genes, and TP53.