ABSTRACT
Advances in genome sequencing have led to a tremendous increase in the discovery of novel missense variants, but evidence for determining clinical significance can be limited or conflicting. Here, we present Learning from Evidence to Assess Pathogenicity (LEAP), a machine learning model that utilizes a variety of feature categories to classify variants, and achieves high performance in multiple genes and different health conditions. Feature categories include functional predictions, splice predictions, population frequencies, conservation scores, protein domain data, and clinical observation data such as personal and family history and covariant information. L2-regularized logistic regression and random forest classification models were trained on missense variants detected and classified during the course of routine clinical testing at Color Genomics (14,226 variants from 24 cancer-related genes and 5,398 variants from 30 cardiovascular-related genes). Using 10-fold cross-validated predictions, the logistic regression model achieved an area under the receiver operating characteristic curve (AUROC) of 97.8% (cancer) and 98.8% (cardiovascular), while the random forest model achieved 98.3% (cancer) and 98.6% (cardiovascular). We demonstrate generalizability to different genes by validating predictions on genes withheld from training (96.8% AUROC). High accuracy and broad applicability make LEAP effective in the clinical setting as a high-throughput quality control layer.
Subject(s)
Genomics/methods , Machine Learning , Models, Genetic , Mutation, Missense , Area Under Curve , Cardiovascular Diseases/genetics , Humans , Logistic Models , Models, Statistical , Neoplasms/genetics , ROC CurveABSTRACT
DNA:RNA hybrids can lead to DNA damage and genome instability. This damage can be prevented by degradation of the RNA in the hybrid by two evolutionarily conserved enzymes, RNase H1 and H2. Indeed, RNase H-deficient cells have increased chromosomal rearrangements. However, the quantitative and spatial contributions of the individual enzymes to hybrid removal have been unclear. Additionally, RNase H2 can remove single ribonucleotides misincorporated into DNA during replication. The relative contribution of DNA:RNA hybrids and misincorporated ribonucleotides to chromosome instability also was uncertain. To address these issues, we studied the frequency and location of loss-of-heterozygosity (LOH) events on chromosome III in Saccharomyces cerevisiae strains that were defective for RNase H1, H2, or both. We showed that RNase H2 plays the major role in preventing chromosome III instability through its hybrid-removal activity. Furthermore, RNase H2 acts pervasively at many hybrids along the chromosome. In contrast, RNase H1 acts to prevent LOH within a small region of chromosome III where the instability is dependent upon two hybrid-prone sequences. This restriction of RNase H1 activity to a subset of hybrids is not the result of its constrained localization, because we found it at hybrids genome-wide. This result suggests that the genome-protection activity of RNase H1 is regulated at a step after hybrid recognition. The global function of RNase H2 and the region-specific function of RNase H1 provide insight into why these enzymes with overlapping hybrid-removal activities have been conserved throughout evolution.
Subject(s)
Chromosomal Instability/genetics , Loss of Heterozygosity/genetics , Ribonuclease H/genetics , Chromosomes, Fungal/genetics , DNA Damage/genetics , DNA Replication/genetics , DNA, Fungal/genetics , RNA, Fungal/genetics , Ribonucleotides/genetics , Saccharomyces cerevisiae/geneticsABSTRACT
BACKGROUND: Next generation sequencing (NGS) has become a common technology for clinical genetic tests. The quality of NGS calls varies widely and is influenced by features like reference sequence characteristics, read depth, and mapping accuracy. With recent advances in NGS technology and software tools, the majority of variants called using NGS alone are in fact accurate and reliable. However, a small subset of difficult-to-call variants that still do require orthogonal confirmation exist. For this reason, many clinical laboratories confirm NGS results using orthogonal technologies such as Sanger sequencing. Here, we report the development of a deterministic machine-learning-based model to differentiate between these two types of variant calls: those that do not require confirmation using an orthogonal technology (high confidence), and those that require additional quality testing (low confidence). This approach allows reliable NGS-based calling in a clinical setting by identifying the few important variant calls that require orthogonal confirmation. RESULTS: We developed and tested the model using a set of 7179 variants identified by a targeted NGS panel and re-tested by Sanger sequencing. The model incorporated several signals of sequence characteristics and call quality to determine if a variant was identified at high or low confidence. The model was tuned to eliminate false positives, defined as variants that were called by NGS but not confirmed by Sanger sequencing. The model achieved very high accuracy: 99.4% (95% confidence interval: +/- 0.03%). It categorized 92.2% (6622/7179) of the variants as high confidence, and 100% of these were confirmed to be present by Sanger sequencing. Among the variants that were categorized as low confidence, defined as NGS calls of low quality that are likely to be artifacts, 92.1% (513/557) were found to be not present by Sanger sequencing. CONCLUSIONS: This work shows that NGS data contains sufficient characteristics for a machine-learning-based model to differentiate low from high confidence variants. Additionally, it reveals the importance of incorporating site-specific features as well as variant call features in such a model.
Subject(s)
High-Throughput Nucleotide Sequencing , Machine Learning , Models, Statistical , Base Sequence , Genetic VariationABSTRACT
The relevance of large copy number variants (CNVs) to hereditary disorders has been long recognized, and population sequencing efforts have chronicled many common structural variants (SVs). However, limited data are available on the clinical contribution of rare germline SVs. Here, a detailed characterization of SVs identified using targeted next-generation sequencing was performed. Across 50 genes associated with hereditary cancer and cardiovascular disorders, a minimum of 828 unique SVs were reported, including 584 fully characterized SVs. Almost 40% of CNVs were <5 kb, with one in three deletions impacting a single exon. Additionally, 36 mid-range deletions/duplications (50 to 250 bp), 21 mobile element insertions, 6 inversions, and 27 complex rearrangements were detected. This data set was used to model SV detection in a bioinformatics pipeline solely relying on read depth, which revealed that genome sequencing (30×) allows detection of 71%, a 500× panel only targeting coding regions 53%, and exome sequencing (100×) <20% of characterized SVs. SVs accounted for 14.1% of all unique pathogenic variants, supporting the importance of SVs in hereditary disorders. Robust SV detection requires an ensemble of variant-calling algorithms that utilize sequencing of intronic regions. These algorithms should use distinct data features representative of each class of mutational mechanism, including recombination between two sequences sharing high similarity, covariants inserted between CNV breakpoints, and complex rearrangements containing inverted sequences.
Subject(s)
Chromosome Breakage , Chromosomes, Human/genetics , Disease/genetics , Genome, Human , Germ-Line Mutation , High-Throughput Nucleotide Sequencing/methods , Introns , Algorithms , HumansABSTRACT
Publicly available genetic databases promote data sharing and fuel scientific discoveries for the prevention, treatment and management of disease. In 2018, we built Color Data, a user-friendly, open access database containing genotypic and self-reported phenotypic information from 50 000 individuals who were sequenced for 30 genes associated with hereditary cancer. In a continued effort to promote access to these types of data, we launched Color Data v2, an updated version of the Color Data database. This new release includes additional clinical genetic testing results from more than 18 000 individuals who were sequenced for 30 genes associated with hereditary cardiovascular conditions as well as polygenic risk scores for breast cancer, coronary artery disease and atrial fibrillation. In addition, we used self-reported phenotypic information to implement the following four clinical risk models: Gail Model for 5-year risk of breast cancer, Claus Model for lifetime risk of breast cancer, simple office-based Framingham Coronary Heart Disease Risk Score for 10-year risk of coronary heart disease and CHARGE-AF simple score for 5-year risk of atrial fibrillation. These new features and capabilities are highlighted through two sample queries in the database. We hope that the broad dissemination of these data will help researchers continue to explore genotype-phenotype correlations and identify novel variants for functional analysis, enabling scientific discoveries in the field of population genomics. Database URL: https://data.color.com/.
Subject(s)
Breast Neoplasms , Genetic Predisposition to Disease , Databases, Factual , Female , Genetic Association Studies , Genotype , HumansABSTRACT
In cascade testing, genetic testing for an identified familial pathogenic variant extends to disease-free relatives to allow genetically targeted disease prevention. We evaluated the results of an online initiative in which carriers of 1 of 30 cancer-associated genes, or their first-degree relatives, could offer low-cost testing to at-risk first-degree relatives. In the first year, 1101 applicants invited 2280 first-degree relatives to undergo genetic testing. Of invited relatives, 47.5% (95% confidence interval [CI] = 45.5 to 49.6%) underwent genetic testing, and 12.0% (95% CI = 9.2 to 14.8%) who tested positive continued the cascade by inviting additional relatives to test. Of tested relatives, 4.9% (95% CI = 3.8 to 6.1%) had a pathogenic variant in a different gene from the known familial one, and 16.8% (95% CI = 14.7 to 18.8%) had a variant of uncertain significance. These results suggest that an online, low-cost program is an effective approach to implementing cascade testing, and that up to 5% of the general population may carry a pathogenic variant in 1 of 30 cancer-associated genes.
Subject(s)
Family , Genetic Carrier Screening/methods , Genetic Predisposition to Disease , Germ-Line Mutation , Neoplasms/diagnosis , Neoplasms/genetics , Online Systems , Humans , PrognosisABSTRACT
Next generation sequencing multi-gene panels have greatly improved the diagnostic yield and cost effectiveness of genetic testing and are rapidly being integrated into the clinic for hereditary cancer risk. With this technology comes a dramatic increase in the volume, type and complexity of data. This invaluable data though is too often buried or inaccessible to researchers, especially to those without strong analytical or programming skills. To effectively share comprehensive, integrated genotypic-phenotypic data, we built Color Data, a publicly available, cloud-based database that supports broad access and data literacy. The database is composed of 50 000 individuals who were sequenced for 30 genes associated with hereditary cancer risk and provides useful information on allele frequency and variant classification, as well as associated phenotypic information such as demographics and personal and family history. Our user-friendly interface allows researchers to easily execute their own queries with filtering, and the results of queries can be shared and/or downloaded. The rapid and broad dissemination of these research results will help increase the value of, and reduce the waste in, scientific resources and data. Furthermore, the database is able to quickly scale and support integration of additional genes and human hereditary conditions. We hope that this database will help researchers and scientists explore genotype-phenotype correlations in hereditary cancer, identify novel variants for functional analysis and enable data-driven drug discovery and development.
Subject(s)
Databases, Genetic , Genetic Variation , Adult , Alleles , BRCA1 Protein/genetics , BRCA2 Protein/genetics , Colorectal Neoplasms, Hereditary Nonpolyposis/genetics , Female , Founder Effect , Genotype , Humans , Jews/genetics , Male , Middle Aged , Phenotype , Search Engine , User-Computer InterfaceABSTRACT
Recent advancements in next-generation sequencing have greatly expanded the use of multi-gene panel testing for hereditary cancer risk. Although genetic testing helps guide clinical diagnosis and management, testing recommendations are based on personal and family history of cancer and ethnicity, and many carriers are being missed. Herein, we report the results from 23,179 individuals who were referred for 30-gene next-generation sequencing panel testing for hereditary cancer risk, independent of current testing guidelines-38.7% of individuals would not have met National Comprehensive Cancer Network criteria for genetic testing. We identified a total of 2811 pathogenic variants in 2698 individuals for an overall pathogenic frequency of 11.6% (9.1%, excluding common low-penetrance alleles). Among individuals of Ashkenazi Jewish descent, three-quarters of pathogenic variants were outside of the three common BRCA1 and BRCA2 founder alleles. Across all ethnic groups, pathogenic variants in BRCA1 and BRCA2 occurred most frequently, but the contribution of pathogenic variants in other genes on the panel varied. Finally, we found that 21.7% of individuals with pathogenic variants in genes with well-established genetic testing recommendations did not meet corresponding National Comprehensive Cancer Network criteria. Taken together, the results indicate that more individuals are at genetic risk for hereditary cancer than are identified by current testing guidelines and/or use of single-gene or single-site testing.
Subject(s)
Biomarkers, Tumor , Genetic Testing , Heterozygote , Neoplastic Syndromes, Hereditary/diagnosis , Neoplastic Syndromes, Hereditary/genetics , Adolescent , Adult , Aged , Aged, 80 and over , Alleles , Female , Gene Frequency , Genetic Predisposition to Disease , Genetic Testing/methods , Humans , Male , Middle Aged , Mutation , Neoplastic Syndromes, Hereditary/mortality , Practice Guidelines as Topic , Prognosis , Young AdultABSTRACT
In Saccharomyces cerevisiae, transcription of the MET regulon, which encodes the proteins involved in the synthesis of the sulfur-containing amino acids methionine and cysteine, is repressed by the presence of either methionine or cysteine in the environment. This repression is accomplished by ubiquitination of the transcription factor Met4, which is carried out by the SCF(Met30) E3 ubiquitin ligase. Mutants defective in MET regulon repression reveal that loss of Cho2, which is required for the methylation of phosphatidylethanolamine to produce phosphatidylcholine, leads to induction of the MET regulon. This induction is due to reduced cysteine synthesis caused by the Cho2 defects, uncovering an important link between phospholipid synthesis and cysteine synthesis. Antimorphic mutants in S-adenosyl-methionine (SAM) synthetase genes also induce the MET regulon. This effect is due, at least in part, to SAM deficiency controlling the MET regulon independently of SAM's contribution to cysteine synthesis. Finally, the Met30 protein is found in two distinct forms whose relative abundance is controlled by the availability of sulfur-containing amino acids. This modification could be involved in the nutritional control of SCF(Met30) activity toward Met4.