Development and validation of a new algorithm for the reclassification of genetic variants identified in the BRCA1 and BRCA2 genes

Abstract

BRCA1 and BRCA2 sequencing analysis detects variants of uncertain clinical significance in approximately 2 % of patients undergoing clinical diagnostic testing in our laboratory. The reclassification of these variants into either a pathogenic or benign clinical interpretation is critical for improved patient management. We developed a statistical variant reclassification tool based on the premise that probands with disease-causing mutations are expected to have more severe personal and family histories than those having benign variants. The algorithm was validated using simulated variants based on approximately 145,000 probands, as well as 286 BRCA1 and 303 BRCA2 true variants. Positive and negative predictive values of ≥99 % were obtained for each gene. Although the history weighting algorithm was not designed to detect alleles of lower penetrance, analysis of the hypomorphic mutations c.5096G>A (p.Arg1699Gln; BRCA1) and c.7878G>C (p.Trp2626Cys; BRCA2) indicated that the history weighting algorithm is able to identify some lower penetrance alleles. The history weighting algorithm is a powerful tool that accurately assigns actionable clinical classifications to variants of uncertain clinical significance. While being developed for reclassification of BRCA1 and BRCA2 variants, the history weighting algorithm is expected to be applicable to other cancer- and non-cancer-related genes.

Assessment of in silico protein sequence analysis in the clinical classification of variants in cancer risk genes

“…Again, algorithm specificities (41.5–59.2%) and PPVs (19.9–27.4%) were low, which is representative of a large number of false positives (Online Resource 2). For example, BRCA1 c.5317A>T (p. Thr1773Ser) was classified as benign by the reference laboratory based on published literature showing that the variant does not affect gene function (Lee et al 2010) and phenotypic evidence from a family history weighting algorithm (Pruss et al 2014) (Table 3). Although the variant was classified as benign by PolyPhen-2, it was classified as pathogenic by Align-GVGD and SIFT.…”

“…• PolyPhen-2 (HumDiv and HumVar)Pathogenic• Align-GVGD, SIFT BRCA2 c.7994A>G (p. Asp2665Gly)BenignPathogenic/Likely Pathogenic• in trans observation with BRCA2 c.4398_4402del (p. Leu1466Phefs*2), which results in premature truncation of the BRCA2 protein at amino acid position 1467 and is classified as pathogenic. The patient lacked symptoms of the recessive phenotype associated with biallelic BRCA2 mutations.• Phenotypic evidence based on family history weighting algorithm (Pruss et al 2014) shows the variant is associated with less severe family history of BRCA2 -associated cancers.• Align-GVGD, PolyPhen-2 (HumDiv and HumVar), SIFT MLH1 c.394G>C (p. Asp132His)BenignPathogenic/Likely Pathogenic• Phenotypic evidence based on family history weighting algorithm (Pruss et al 2014) shows the variant is associated with less severe family history of MLH1 -associated cancers.• in trans observation with the pathogenic MLH1 variant, c.1731G>A (p. Ser577Ser), which is a silent variant that causes skipping of exon 15 (Pagenstecher et al 2006) and has been described in multiple HNPCC families (Kohonen-Corish et al 1996; Pagenstecher et al 2006; Wu et al 1997). The patient lacked symptoms of the recessive phenotype associated with biallelic MLH1 mutations.• Align-GVGD, PolyPhen-2 (HumDiv and HumVar), SIFT, MAPP-MMR …”

“…These variants were classified as pathogenic or benign based on multiple, independent lines of evidence that did not include sequence conservation (Eggington et al 2013; Richards et al 2015). This includes statistical methods (Pruss et al 2014), biochemical/biophysical assays in the literature, population allele frequency thresholds, and co-occurrences with mutations in the same gene or pathway. Variants whose pathogenicity may arise from a mechanism not purely related to the amino acid substitution (variants that impact mRNA splicing, translation from the initiating 5′ AUG start codon, silent mutations) were excluded.…”

Beyond DNA: An Integrated and Functional Approach for Classifying Germline Variants in Breast Cancer Genes

“…For example, when assessing a TP53 variant, the phenotype is considered strong if the patient meets classic Li-Fraumeni syndrome criteria: a proband with sarcoma diagnosed before 45 years, a first-degree relative with any cancer before 45 years, and a second-degree relative with any cancer before age 45 years or a sarcoma at any age [15]. For common diseases and moderate penetrance genes Bayesian analyses that require larger phenotype data sets are used [16]. Historically, in vitro studies were predominantly found in the published literature.…”

Look Before You Leap

“…This classification incorporates multiple factors, including previous reports about the variant or data from internal or external databases. 12,13 Because many variants are novel and have never been reported in the medical literature and because of limitations of the existing literature, classifying genetic variants as disease-causing is subject to uncertainty and possible misclassification. 14 Differences in the available knowledge base can lead to interlaboratory discrepancies.…”