Evaluation of <Emphasis Type="Italic">in silico</Emphasis> algorithms for use with ACMG/AMP clinical variant interpretation guidelines

Background

The American College of Medical Genetics and American College of Pathologists (ACMG/AMP) variant classification guidelines for clinical reporting are widely used in diagnostic laboratories for variant interpretation. The ACMG/AMP guidelines recommend complete concordance of predictions among all in silico algorithms used without specifying the number or types of algorithms. The subjective nature of this recommendation contributes to discordance of variant classification among clinical laboratories and prevents definitive classification of variants.

Results

Using 14,819 benign or pathogenic missense variants from the ClinVar database, we compared performance of 25 algorithms across datasets differing in distinct biological and technical variables. There was wide variability in concordance among different combinations of algorithms with particularly low concordance for benign variants. We also identify a previously unreported source of error in variant interpretation (false concordance) where concordant in silico predictions are opposite to the evidence provided by other sources. We identified recently developed algorithms with high predictive power and robust to variables such as disease mechanism, gene constraint, and mode of inheritance, although poorer performing algorithms are more frequently used based on review of the clinical genetics literature (2011–2017).

Conclusions

Our analyses identify algorithms with high performance characteristics independent of underlying disease mechanisms. We describe combinations of algorithms with increased concordance that should improve in silico algorithm usage during assessment of clinically relevant variants using the ACMG/AMP guidelines.

     

Trans-activation-based risk assessment of <Emphasis Type="Italic">BRCA1</Emphasis> BRCT variants with unknown clinical significance

Background

Deleterious variants in the tumour suppressor BRCA1 are known to cause hereditary breast and ovarian cancer syndrome (HBOC). Missense variants in BRCA1 pose a challenge in clinical care, as their effect on protein functionality often remains unknown. Many of the pathogenic missense variants found in BRCA1 are located in the BRCA1 C-terminal (BRCT) domains, domains that are known to be vital for key functions such as homologous recombination repair, protein-protein interactions and trans-activation (TA). We investigated the TA activity of 12 BRCA1 variants of unknown clinical significance (VUSs) located in the BRCT domains to aid in the classification of these variants.

Results

Twelve BRCA1 VUSs were investigated using a modified version of the dual luciferase TA activity assay (TA assay) that yielded increased sensitivity and sample throughput. Variants were classified according to American College of Medical Genetics and Genomics (ACMG) criteria using TA assay results and available data. In combining our TA-assay results and available data, in accordance with the ACMG guidelines for variant classification, we proposed the following variant classifications: c.5100A>G, c.5326C>T, c.5348T>C and c.5477A>T as likely benign (class 2) variants. c.5075A>C, c.5116G>A and c.5513T>G were likely pathogenic (class 4), whereas c.5096G>A likely represents a likely pathogenic variant with moderate penetrance. Variants c.5123C>T, c.5125G>A, c.5131A>C and c.5504G>A remained classified as VUSs (class 3).

Conclusions

The modified TA assay provides efficient risk assessment of rare missense variants found in the BRCA1 BRCT-domains. We also report that increased post-transfection incubation time yielded a significant increase in TA assay sensitivity.

     

Copy number variation (CNV) identification,interpretation, and database from Brazilian patients

Copy number variations (CNVs) constitute an important class of variation in the human genome and the interpretation of their pathogenicity considering different frequencies across populations is still a challenge for geneticists. Since the CNV databases are predominantly composed of European and non-admixed individuals, and Brazilian genetic constitution is admixed and ethnically diverse, diagnostic screenings on Brazilian variants are greatly difficulted by the lack of populational references. We analyzed a clinical sample of 268 Brazilian individuals, including patients with neurodevelopment disorders and/or congenital malformations. The pathogenicity of CNVs was classified according to their gene content and overlap with known benign and pathogenic variants. A total of 1,504 autosomal CNVs (1,207 gains and 297 losses) were classified as benign (92.9%), likely benign (1.6%), VUS (2.6%), likely pathogenic (0.2%) and pathogenic (2.7%). Some of the CNVs were recurrent and with frequency increased in our sample, when compared to populational open resources of structural variants: 14q32.33, 22q11.22, 1q21.1, and 1p36.32 gains. Thus, these highly recurrent CNVs classified as likely benign or VUS were considered non-pathogenic in our Brazilian sample. This study shows the relevance of introducing CNV data from diverse cohorts to improve on the interpretation of clinical impact of genomic variations.     

Burden of rare sarcomere gene variants in the framingham and jackson heart study cohorts

Rare sarcomere protein variants cause dominant hypertrophic and dilated cardiomyopathies. To evaluate whether allelic variants in eight sarcomere genes are associated with cardiac morphology and function in the community, we sequenced 3,600 individuals from the Framingham Heart Study (FHS) and Jackson Heart Study (JHS) cohorts. Out of the total, 11.2% of individuals had one or more rare nonsynonymous sarcomere variants. The prevalence of likely pathogenic sarcomere variants was 0.6%, twice the previous estimates; however, only four of the 22 individuals had clinical manifestations of hypertrophic cardiomyopathy. Rare sarcomere variants were associated with an increased risk for adverse cardiovascular events (hazard ratio: 2.3) in the FHS cohort, suggesting that cardiovascular risk assessment in the general population can benefit from rare variant analysis.     

A scalable Drosophila assay for clinical interpretation of human PTEN variants in suppression of PI3K/AKT induced cellular proliferation

Gene variant discovery is becoming routine, but it remains difficult to usefully interpret the functional consequence or disease relevance of most variants. To fill this interpretation gap, experimental assays of variant function are becoming common place. Yet, it remains challenging to make these assays reproducible, scalable to high numbers of variants, and capable of assessing defined gene-disease mechanism for clinical interpretation aligned to the ClinGen Sequence Variant Interpretation (SVI) Working Group guidelines for ‘well-established assays’. Drosophila melanogaster offers great potential as an assay platform, but was untested for high numbers of human variants adherent to these guidelines. Here, we wished to test the utility of Drosophila as a platform for scalable well-established assays. We took a genetic interaction approach to test the function of ~100 human PTEN variants in cancer-relevant suppression of PI3K/AKT signaling in cellular growth and proliferation. We validated the assay using biochemically characterized PTEN mutants as well as 23 total known pathogenic and benign PTEN variants, all of which the assay correctly assigned into predicted functional categories. Additionally, function calls for these variants correlated very well with our recent published data from a human cell line. Finally, using these pathogenic and benign variants to calibrate the assay, we could set readout thresholds for clinical interpretation of the pathogenicity of 70 other PTEN variants. Overall, we demonstrate that Drosophila offers a powerful assay platform for clinical variant interpretation, that can be used in conjunction with other well-established assays, to increase confidence in the accurate assessment of variant function and pathogenicity.     

Identification of Medically Actionable Secondary Findings in the 1000 Genomes

The American College of Medical Genetics and Genomics (ACMG) recommends that clinical sequencing laboratories return secondary findings in 56 genes associated with medically actionable conditions. Our goal was to apply a systematic, stringent approach consistent with clinical standards to estimate the prevalence of pathogenic variants associated with such conditions using a diverse sequencing reference sample. Candidate variants in the 56 ACMG genes were selected from Phase 1 of the 1000 Genomes dataset, which contains sequencing information on 1,092 unrelated individuals from across the world. These variants were filtered using the Human Gene Mutation Database (HGMD) Professional version and defined parameters, appraised through literature review, and examined by a clinical laboratory specialist and expert physician. Over 70,000 genetic variants were extracted from the 56 genes, and filtering identified 237 variants annotated as disease causing by HGMD Professional. Literature review and expert evaluation determined that 7 of these variants were pathogenic or likely pathogenic. Furthermore, 5 additional truncating variants not listed as disease causing in HGMD Professional were identified as likely pathogenic. These 12 secondary findings are associated with diseases that could inform medical follow-up, including cancer predisposition syndromes, cardiac conditions, and familial hypercholesterolemia. The majority of the identified medically actionable findings were in individuals from the European (5/379) and Americas (4/181) ancestry groups, with fewer findings in Asian (2/286) and African (1/246) ancestry groups. Our results suggest that medically relevant secondary findings can be identified in approximately 1% (12/1092) of individuals in a diverse reference sample. As clinical sequencing laboratories continue to implement the ACMG recommendations, our results highlight that at least a small number of potentially important secondary findings can be selected for return. Our results also confirm that understudied populations will not reap proportionate benefits of genomic medicine, highlighting the need for continued research efforts on genetic diseases in these populations.     

Germline variants in oculocutaneous albinism genes and predisposition to familial cutaneous melanoma

Approximately 1%–2% of cutaneous melanoma (CM) is classified as strongly familial. We sought to investigate unexplained CM predisposition in families negative for the known susceptibility genes using next‐generation sequencing of affected individuals. Segregation of germline variants of interest within families was assessed by Sanger sequencing. Several heterozygous variants in oculocutaneous albinism (OCA) genes: TYR, OCA2, TYRP1 and SLC45A2, were present in our CM cohort. OCA is a group of autosomal recessive genetic disorders, resulting in pigmentation defects of the eyes, hair and skin. Missense variants classified as pathogenic for OCA were present in multiple families and some fully segregated with CM. The functionally compromised TYR p.T373K variant was present in three unrelated families. In OCA2, known pathogenic variants: p.V443I and p.N489D, were present in three families and one family, respectively. We identified a likely pathogenic SLC45A2 frameshift variant that fully segregated with CM in a family of four cases. Another four‐case family harboured cosegregating variants (p.A24T and p.R153C) of uncertain functional significance in TYRP1. We conclude that rare, heterozygous variants in OCA genes confer moderate risk for CM.     

High resolution array in the clinical approach to chromosomal phenotypes

Array genomic hybridization (AGH) has recently been implemented as a diagnostic tool for the detection of submicroscopic copy number variants (CNVs) in patients with developmental disorders. However, there is no consensus regarding the choice of the platform, the minimal resolution needed and systematic interpretation of CNVs. We report our experience in the clinical diagnostic use of high resolution AGH up to 100 kb on 131 patients with chromosomal phenotypes but previously normal karyotype. We evaluated the usefulness in our clinics and laboratories by the detection rate of causal CNVs and CNVs of unknown clinical significance and to what extent their interpretation would challenge the systematic use of high-resolution arrays in clinical application. Prioritizing phenotype-genotype correlation in our interpretation strategy to criteria previously described, we identified 33 (25.2%) potentially pathogenic aberrations. 16 aberrations were confirmed pathogenic (16.4% syndromic, 8.5% non-syndromic patients); 9 were new and individual aberrations, 3 of them were pathogenic although inherited and one is as small as approx 200 kb. 13 of 16 further CNVs of unknown significance were classified likely benign, for 3 the significance remained unclear. High resolution array allows the detection of up to 12.2% of pathogenic aberrations in a diagnostic clinical setting. Although the majority of aberrations are larger, the detection of small causal aberrations may be relevant for family counseling. The number of remaining unclear CNVs is limited. Careful phenotype-genotype correlations of the individual CNVs and clinical features are challenging but remain a hallmark for CNV interpretation.     

The impact of post-alignment processing procedures on whole-exome sequencing data

The use of post-alignment procedures has been suggested to prevent the identification of false-positives in massive DNA sequencing data. Insertions and deletions are most likely to be misinterpreted by variant calling algorithms. Using known genetic variants as references for post-processing pipelines can minimize mismatches. They allow reads to be correctly realigned and recalibrated, resulting in more parsimonious variant calling. In this work, we aim to investigate the impact of using different sets of common variants as references to facilitate variant calling from whole-exome sequencing data. We selected reference variants from common insertions and deletions available within the 1K Genomes project data and from databases from the Latin American Database of Genetic Variation (LatinGen). We used the Genome Analysis Toolkit to perform post-processing procedures like local realignment, quality recalibration procedures, and variant calling in whole exome samples. We identified an increased number of variants from the call set for all groups when no post-processing procedure was performed. We found that there was a higher concordance rate between variants called using 1K Genomes and LatinGen. Therefore, we believe that the increased number of rare variants identified in the analysis without realignment or quality recalibration indicated that they were likely false-positives.     

RAD51C germline mutations in breast and ovarian cancer cases from high-risk families

BRCA1 and BRCA2 are the most well-known breast cancer susceptibility genes. Additional genes involved in DNA repair have been identified as predisposing to breast cancer. One such gene, RAD51C, is essential for homologous recombination repair. Several likely pathogenic RAD51C mutations have been identified in BRCA1- and BRCA2-negative breast and ovarian cancer families. We performed complete sequencing of RAD51C in germline DNA of 286 female breast and/or ovarian cancer cases with a family history of breast and ovarian cancers, who had previously tested negative for mutations in BRCA1 and BRCA2. We screened 133 breast cancer cases, 119 ovarian cancer cases, and 34 with both breast and ovarian cancers. Fifteen DNA sequence variants were identified; including four intronic, one 5' UTR, one promoter, three synonymous, and six non-synonymous variants. None were truncating. The in-silico SIFT and Polyphen programs were used to predict possible pathogenicity of the six non-synonomous variants based on sequence conservation. G153D and T287A were predicted to be likely pathogenic. Two additional variants, A126T and R214C alter amino acids in important domains of the protein such that they could be pathogenic. Two-hybrid screening and immunoblot analyses were performed to assess the functionality of these four non-synonomous variants in yeast. The RAD51C-G153D protein displayed no detectable interaction with either XRCC3 or RAD51B, and RAD51C-R214C displayed significantly decreased interaction with both XRCC3 and RAD51B (p<0.001). Immunoblots of RAD51C-Gal4 activation domain fusion peptides showed protein levels of RAD51C-G153D and RAD51C-R214C that were 50% and 60% of the wild-type, respectively. Based on these data, the RAD51C-G153D variant is likely to be pathogenic, while the RAD51C- R214C variant is hypomorphic of uncertain pathogenicity. These results provide further support that RAD51C is a rare breast and ovarian cancer susceptibility gene.     

A novel XPC pathogenic variant detected in archival material from a patient diagnosed with Xeroderma Pigmentosum: a case report and review of the genetic variants reported in XPC

The disease Xeroderma Pigmentosum (XP) is genetically heterogeneous and defined by pathogenic variants (formerly termed mutations) in any of eight different genes. Pathogenic variants in the XPC gene are the most commonly observed in US patients. Moreover, pathogenic variants in just four of the genes, XPA, XPC, XPD/ERCC2 and XPV/POLH account for 91% of all XP cases worldwide. In the current study, we describe the clinical, histopathologic, molecular genetic, and pathophysiological features of a 19-year-old female patient clinically diagnosed with XP as an infant. Analysis of archival material reveals a novel variation of a 13 base pair deletion in XPC exon 14 and a previously reported A>C missense pathogenic variant in the proximal splice site for XPC exon 6. Both variations induce frameshifts most likely leading to a truncated XPC protein product. Quantitative RT-PCR also revealed reduced mRNA levels in the archived specimen. Analysis of the XPA, XPD/ERCC2 and XPV/POLH genes in the current specimen failed to reveal pathologic variants. All previously reported pathogenic variants, polymorphisms and known amino acid changes for the XPC gene are compiled and described in the current nomenclature. Given the relative ease of screening for genetic variation and the potential role for such variation in human disease, a proposal for screening appropriate archival materials for alterations in the four most prevalent XP genes is presented.     

Utilizing Ethnic-Specific Differences in Minor Allele Frequency to Recategorize Reported Pathogenic Deafness Variants

Ethnic-specific differences in minor allele frequency impact variant categorization for genetic screening of nonsyndromic hearing loss (NSHL) and other genetic disorders. We sought to evaluate all previously reported pathogenic NSHL variants in the context of a large number of controls from ethnically distinct populations sequenced with orthogonal massively parallel sequencing methods. We used HGMD, ClinVar, and dbSNP to generate a comprehensive list of reported pathogenic NSHL variants and re-evaluated these variants in the context of 8,595 individuals from 12 populations and 6 ethnically distinct major human evolutionary phylogenetic groups from three sources (Exome Variant Server, 1000 Genomes project, and a control set of individuals created for this study, the OtoDB). Of the 2,197 reported pathogenic deafness variants, 325 (14.8%) were present in at least one of the 8,595 controls, indicating a minor allele frequency (MAF) >0.00006. MAFs ranged as high as 0.72, a level incompatible with pathogenicity for a fully penetrant disease like NSHL. Based on these data, we established MAF thresholds of 0.005 for autosomal-recessive variants (excluding specific variants in GJB2) and 0.0005 for autosomal-dominant variants. Using these thresholds, we recategorized 93 (4.2%) of reported pathogenic variants as benign. Our data show that evaluation of reported pathogenic deafness variants using variant MAFs from multiple distinct ethnicities and sequenced by orthogonal methods provides a powerful filter for determining pathogenicity. The proposed MAF thresholds will facilitate clinical interpretation of variants identified in genetic testing for NSHL. All data are publicly available to facilitate interpretation of genetic variants causing deafness.     

Actionable,Pathogenic Incidental Findings in 1,000 Participants’ Exomes

The incorporation of genomics into medicine is stimulating interest on the return of incidental findings (IFs) from exome and genome sequencing. However, no large-scale study has yet estimated the number of expected actionable findings per individual; therefore, we classified actionable pathogenic single-nucleotide variants in 500 European- and 500 African-descent participants randomly selected from the National Heart, Lung, and Blood Institute Exome Sequencing Project. The 1,000 individuals were screened for variants in 114 genes selected by an expert panel for their association with medically actionable genetic conditions possibly undiagnosed in adults. Among the 1,000 participants, 585 instances of 239 unique variants were identified as disease causing in the Human Gene Mutation Database (HGMD). The primary literature supporting the variants’ pathogenicity was reviewed. Of the identified IFs, only 16 unique autosomal-dominant variants in 17 individuals were assessed to be pathogenic or likely pathogenic, and one participant had two pathogenic variants for an autosomal-recessive disease. Furthermore, one pathogenic and four likely pathogenic variants not listed as disease causing in HGMD were identified. These data can provide an estimate of the frequency (∼3.4% for European descent and ∼1.2% for African descent) of the high-penetrance actionable pathogenic or likely pathogenic variants in adults. The 23 participants with pathogenic or likely pathogenic variants were disproportionately of European (17) versus African (6) descent. The process of classifying these variants underscores the need for a more comprehensive and diverse centralized resource to provide curated information on pathogenicity for clinical use to minimize health disparities in genomic medicine.     

Targeted next-generation sequencing to diagnose disorders of HDL cholesterol

A low level of HDL cholesterol (HDL-C) is a common clinical scenario and an important marker for increased cardiovascular risk. Many patients with very low or very high HDL-C have a rare mutation in one of several genes, but identification of the molecular abnormality in patients with extreme HDL-C is rarely performed in clinical practice. We investigated the accuracy and diagnostic yield of a targeted next-generation sequencing (NGS) assay for extreme levels of HDL-C. We developed a targeted NGS panel to capture the exons, intron/exon boundaries, and untranslated regions of 26 genes with highly penetrant effects on plasma lipid levels. We sequenced 141 patients with extreme HDL-C levels and prioritized variants in accordance with medical genetics guidelines. We identified 35 pathogenic and probably pathogenic variants in HDL genes, including 21 novel variants, and performed functional validation on a subset of these. Overall, a molecular diagnosis was established in 35.9% of patients with low HDL-C and 5.2% with high HDL-C, and all prioritized variants identified by NGS were confirmed by Sanger sequencing. Our results suggest that a molecular diagnosis can be identified in a substantial proportion of patients with low HDL-C using targeted NGS.     

Chromosome patterns of nontransformed variants from chemically transformed Balb-3T3 cells

Nine variant cell lines isolated from cloned 7,12-dimethylbenz(a) -ahthracene transformed Balb/3T3 mouse cells by treatment with FUdR had growth parameters closely resembling nontransformed cells. Chromosome analysis of the variant lines demonstrated that six variants had a diminished number and three variants had an increased number of chromosomes compared to the parental transformed cell line. All variants had unique marker chromosomes not present in the parental transformed Balb/3T3 cells. The distribution of marker chromosomes and heterochromatin suggested that the initial event in variant formation was a reduction in chromosome number with a subsequent polyploidization of the reduced chromosome complement.     

Decoding the effects of synonymous variants

Synonymous single nucleotide variants (sSNVs) are common in the human genome but are often overlooked. However, sSNVs can have significant biological impact and may lead to disease. Existing computational methods for evaluating the effect of sSNVs suffer from the lack of gold-standard training/evaluation data and exhibit over-reliance on sequence conservation signals. We developed synVep (synonymous Variant effect predictor), a machine learning-based method that overcomes both of these limitations. Our training data was a combination of variants reported by gnomAD (observed) and those unreported, but possible in the human genome (generated). We used positive-unlabeled learning to purify the generated variant set of any likely unobservable variants. We then trained two sequential extreme gradient boosting models to identify subsets of the remaining variants putatively enriched and depleted in effect. Our method attained 90% precision/recall on a previously unseen set of variants. Furthermore, although synVep does not explicitly use conservation, its scores correlated with evolutionary distances between orthologs in cross-species variation analysis. synVep was also able to differentiate pathogenic vs. benign variants, as well as splice-site disrupting variants (SDV) vs. non-SDVs. Thus, synVep provides an important improvement in annotation of sSNVs, allowing users to focus on variants that most likely harbor effects.     

Cytokines and Tumor Metastasis Gene Variants in Oral Cancer and Precancer in Puerto Rico

Objectives

A cross-sectional epidemiological study explored genetic susceptibility to oral precancer and cancer in Puerto Rico (PR).

Materials and Methods

Three hundred three individuals with a benign oral condition, oral precancer (oral epithelial hyperplasia/hyperkeratosis, oral epithelial dysplasia), or oral squamous cell carcinoma (SCCA) were identified via PR pathology laboratories. A standardized, structured questionnaire obtained information on epidemiological variables; buccal cells were collected for genetic analysis. Genotyping was performed using Taqman® assays. Allelic frequencies of single nucleotide polymorphisms (SNPs) were evaluated in cytokine genes and genes influencing tumor metastasis. Risk estimates for a diagnosis of oral precancer or SCCA while having a variant allele were generated using logistic regression. Adjusted models controlled for age, gender, ancestry, education, smoking and alcohol consumption.

Results

Relative to persons with a benign oral lesion, individuals with homozygous recessive allelic variants of tumor necrosis factor (TNF-α) −238 A/G SNP had a reduced odds of having an oral precancer (OR_adjusted = 0.15; 95% CI 0.03–0.70). The transforming growth factor beta-1 (TGFβ-1 −509 C/T) polymorphism was inversely associated with having an oral SCCA among persons homozygous for the recessive variant (OR_crude = 0.27; 95% CI 0.09–0.79). The matrix metalloproteinase gene (MMP-1) variant, rs5854, was associated with oral SCCA; participants with even one variant allele were more likely to have oral SCCA (OR_adjusted = 2.62, 95% CI 1.05–6.53) compared to people with ancestral alleles.

Conclusion

Our exploratory analyses suggest that genetic alterations in immune system genes and genes with metastatic potential are associated with oral precancer and SCCA risk in PR.     

Annotating DNA Variants Is the Next Major Goal for Human Genetics

Clinical genetic testing has undergone a dramatic transformation in the past two decades. Diagnostic laboratories that previously tested for well-established disease-causing DNA variants in a handful of genes have evolved into sequencing factories identifying thousands of variants of known and unknown medical consequence. Sorting out what does and does not cause disease in our genomes is the next great challenge in making genetics a central feature of healthcare. I propose that closing the gap in our ability to interpret variation responsible for Mendelian disorders provides a grand and unprecedented opportunity for geneticists. Human geneticists are well placed to coordinate a systematic evaluation of variants in collaboration with basic scientists and clinicians. Sharing of knowledge, data, methods, and tools will aid both researchers and healthcare workers in achieving their common goal of defining the pathogenic potential of variants. Generation of variant annotations will inform genetic testing and will deepen our understanding of gene and protein function, thereby aiding the search for molecular targeted therapies.