Gene-based calibration of high-throughput functional assays for clinical variant classification
Zeiberg, D.; Tejura, M.; McEwen, A. E.; Fayer, S.; Pejaver, V.; Rubin, A. F.; Starita, L. M.; Fowler, D. M.; O'Donnell-Luria, A.; Radivojac, P.
Show abstract
High-throughput assays measure a broad range of variant effects on gene function and hold promise for supporting genomic medicine. Current clinical guidelines for rare Mendelian diseases rely on establishing gene-specific score thresholds for each assay that separate pathogenic from benign variants. This introduces inconsistencies and subjectivity, ultimately lacking the rigor of calibration; i.e., mapping a variant score to a probability of pathogenicity. To address this problem, we introduce a semi-supervised framework for calibrating experimental assay data and propose Experimental score CALIBRator (ExCALIBR), a method that jointly models pathogenic, benign, synonymous, and population variants using skew normal mixtures to produce variant-specific probabilities of pathogenicity. Evaluated across 80 datasets from 39 genes, all meeting fit quality criteria, ExCALIBR substantially outperformed existing field standards and was further validated on the All of Us biobank data. Our results demonstrate that calibrated experimental assays generate indispensable evidence that will dramatically reduce variants of uncertain significance.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- STRling: a k-mer counting approach that detects short tandem repeat expansions at known and novel loci 95%
- Variant effect predictor correlation with functional assays is reflective of clinical classification performance 95%
- MTSplice predicts effects of genetic variants on tissue-specific splicing 95%
Similar papers in this journal
- MetaRNN: Differentiating Rare Pathogenic and Rare Benign Missense SNVs and InDels Using Deep Learning 94%
- Genome-wide prediction of pathogenic gain- and loss-of-function variants from ensemble learning of diverse feature set 94%
- Constructing germline research cohorts from the discarded reads of clinical tumor sequences 94%
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- Efficient and Flexible Integration of Variant Characteristics in Rare Variant Association Studies Using Integrated Nested Laplace Approximation 95%
- ForestQC: quality control on genetic variants from next-generation sequencing data using random forest 94%
- Improving the coverage of credible sets in Bayesian genetic fine-mapping 93%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.