Systematic misclassification of missense variants in BRCA1 and BRCA2 "coldspots"

Abstract

Purpose: Guidelines for variant interpretation incorporate variant hotspots in critical functional domains as evidence for pathogenicity (e.g., PM1 and PP2), but do not use "coldspots," that is, regions without essential functions that tolerate variation, as evidence a variant is benign. To improve variant classification we evaluated BRCA1 and BRCA2 missense variants reported in ClinVar to identify regions where pathogenic missenses are extremely infrequent, defined as coldspots.

Methods: We used Bayesian approaches to model variant classification in these regions.

Results: BRCA1 exon 11 (~60% of the coding sequence), and BRCA2 exons 10 and 11 (~65% of the coding sequence), are coldspots. Of 89 pathogenic (P) or likely pathogenic (LP) missense variants in BRCA1, none are in exon 11 (odds <0.01, 95% confidence interval [CI] 0.0-0.01). Of 34 P or LP missense variants in BRCA2, none are in exons 10-11 (odds <0.01, 95% CI 0.0-0.01). More than half of reported missense variants of uncertain significance (VUS) in BRCA1 and BRCA2 are in coldspots (3115/5301 = 58.8%). Reclassifying these 3115 VUS as likely benign would substantially improve variant classification.

Conclusion: In BRCA1 and BRCA2 coldspots, missense variants are very unlikely to be pathogenic. Classification schemes that incorporate coldspots can reduce the number of VUS and mitigate risks from reporting benign variation as VUS.

Keywords: ACMG; BRCA1; VUS; coldspot; variant classification.

Fig. 1. Distributions of missense variants in BRCA1 and BRCA2.

Missense variants are indicated by lollipops. For each gene, distributions of pathogenic and likely pathogenic (P/LP) missense variants are shown in orange above the gene and of variants of uncertain signficance (VUS) in gray below the gene. BRCA1 exon 11 and BRCA2 exons 10 and 11 harbor no P or LP missense variants and are defined as coldspots. We suggest reclassifying the 3115 VUS in these coldspots as likely benign.