The Frequency of Discordant Variant Classification in the Human Gene Mutation Database: A Comparison of the American College of Medical Genetics and Genomics Guidelines and ClinVar

Objective Discordant variant classifications among public databases is one of the well-documented limitations when interpreting the pathogenicity of variants. The aim of this study is to investigate the level of germline variant misannotation from the Human Gene Mutation Database (HGMD) and the annotation concordance between databases. Methods We used a total of 188,106 classified variants (disease-causing mutations [n = 179,454] and polymorphisms [n = 8652]) in 6466 genes from the HGMD. All variants were reanalyzed based on the American College of Medical Genetics and Genomics (ACMG) guidelines and compared to ClinVar database variants. Results When variants were classified based on the ACMG guidelines, misclassification was observed in 3.47% (2289/65,896) of variants. The overall concordance between HGMD and ClinVar was 97.62% (52,499/53,780) of variants studied. Conclusion Variants in databases must be used with caution when variant pathogenicity is interpreted. This study reveals the frequency of misannotation of the HGMD variants and annotation concordance between databases in depth.

The Human Gene Mutation Database (HGMD®): optimizing its use in a clinical diagnostic or research setting

The Human Gene Mutation Database (HGMD®) constitutes a comprehensive collection of published germline mutations in nuclear genes that are thought to underlie, or are closely associated with human inherited disease. At the time of writing (June 2020), the database contains in excess of 289,000 different gene lesions identified in over 11,100 genes manually curated from 72,987 articles published in over 3100 peer-reviewed journals. There are primarily two main groups of users who utilise HGMD on a regular basis; research scientists and clinical diagnosticians. This review aims to highlight how to make the most out of HGMD data in each setting.

Abstract 19585: Overestimation of Functional Cardiomyopathy Disease Variants in the Human Gene Mutation Database

Introduction: Dilated cardiomyopathy (DM), hypertrophic cardiomyopathy (HCM), and arryhmogenic right ventricular dysplasia (ARVD) have been associated with germline genetic variants. However, uncertainty exists regarding the functional impact of specific variants described in the literature. Hypothesis: We hypothesized that a substantial fraction of presumed casual variants would be identified as non-disease causing in a normal patient population. Methods: Missense variants associated with three main subtypes of cardiomyopathy were extracted from Human Gene Mutation Database Professional (HGMD). To assess the prevalence and pathogenicity of these potential disease-causing alleles, the population frequency of each variant was screened in a recently published large-scale exome database of over 60,000 whole exomes (ExAC). In addition, PolyPhen-2 was used to predict the functional impact of amino acid substitution for each missense variant. Results: In the HGMD, 1,405 missense variants were associated with cardiomyopathy, of which 25% were found in 2 or more samples in the ExAC database. Specifically 121 of 422 (28.7%) DM associated variants, 201 variants of 900 (22.3%) HCM associated variant, and 40 of 83 ARVD (48.2%) associated variants were observed in the ExAC database. Polyphen predicted 331 (23.9%) of variants as benign, 290 (20.6%) as possibly damaging, and 784 (55.8%) as probably damaging. Table 1 presents disease specific Polyphen results. Conclusions: We observed that a large proportion (25-45%) of cardiomyopathy-associated missense variants predicted as being damaging were indistinguishable from the background in ExAC. Using stringent cutoffs derived from these observations, we estimated more than 50% of previously associated cardiomyopathy variants may be non-functional or non-monogenic causes of cardiomyopathy.