A high-resolution map of human RNA translation.

Translated small open reading frames (smORFs) can have important regulatory roles and encode microproteins, yet their genome-wide identification has been challenging. We determined the ribosome locations across six primary human cell types and five tissues and detected 7,767 smORFs with translational profiles matching those of known proteins. The human genome was found to contain highly cell-type- and tissue-specific smORFs and a subset that encodes highly conserved amino acid sequences. Changes in the translational efficiency of upstream-encoded smORFs (uORFs) and the corresponding main ORFs predominantly occur in the same direction. Integration with 456 mass-spectrometry datasets confirms the presence of 603 small peptides at the protein level in humans and provides insights into the subcellular localization of these small proteins. This study provides a comprehensive atlas of high-confidence translated smORFs derived from primary human cells and tissues in order to provide a more complete understanding of the translated human genome.

Ribosome decision graphs for the representation of eukaryotic RNA translation complexity.

The application of ribosome profiling has revealed an unexpected abundance of translation in addition to that responsible for the synthesis of previously annotated protein-coding regions. Multiple short sequences have been found to be translated within single RNA molecules, within both annotated protein-coding and noncoding regions. The biological significance of this translation is a matter of intensive investigation. However, current schematic or annotation-based representations of mRNA translation generally do not account for the apparent multitude of translated regions within the same molecules. They also do not take into account the stochasticity of the process that allows alternative translations of the same RNA molecules by different ribosomes. There is a need for formal representations of mRNA complexity that would enable the analysis of quantitative information on translation and more accurate models for predicting the phenotypic effects of genetic variants affecting translation. To address this, we developed a conceptually novel abstraction that we term ribosome decision graphs (RDGs). RDGs represent translation as multiple ribosome paths through untranslated and translated mRNA segments. We termed the latter "translons." Nondeterministic events, such as initiation, reinitiation, selenocysteine insertion, or ribosomal frameshifting, are then represented as branching points. This representation allows for an adequate representation of eukaryotic translation complexity and focuses on locations critical for translation regulation. We show how RDGs can be used for depicting translated regions and for analyzing genetic variation and quantitative genome-wide data on translation for characterization of regulatory modulators of translation.

2022 Curt Stern Award introduction: Heidi Rehm.

This article is based on the address given by the author at the 2022 meeting of the American Society of Human Genetics (ASHG) in Los Angeles, California. The video of the original address can be found at the ASHG website.

Evaluating drug targets through human loss-of-function genetic variation.

Naturally occurring human genetic variants that are predicted to inactivate protein-coding genes provide an in vivo model of human gene inactivation that complements knockout studies in cells and model organisms. Here we report three key findings regarding the assessment of candidate drug targets using human loss-of-function variants. First, even essential genes, in which loss-of-function variants are not tolerated, can be highly successful as targets of inhibitory drugs. Second, in most genes, loss-of-function variants are sufficiently rare that genotype-based ascertainment of homozygous or compound heterozygous 'knockout' humans will await sample sizes that are approximately 1,000 times those presently available, unless recruitment focuses on consanguineous individuals. Third, automated variant annotation and filtering are powerful, but manual curation remains crucial for removing artefacts, and is a prerequisite for recall-by-genotype efforts. Our results provide a roadmap for human knockout studies and should guide the interpretation of loss-of-function variants in drug development.

The mutational constraint spectrum quantified from variation in 141,456 humans.

Genetic variants that inactivate protein-coding genes are a powerful source of information about the phenotypic consequences of gene disruption: genes that are crucial for the function of an organism will be depleted of such variants in natural populations, whereas non-essential genes will tolerate their accumulation. However, predicted loss-of-function variants are enriched for annotation errors, and tend to be found at extremely low frequencies, so their analysis requires careful variant annotation and very large sample sizes1. Here we describe the aggregation of 125,748 exomes and 15,708 genomes from human sequencing studies into the Genome Aggregation Database (gnomAD). We identify 443,769 high-confidence predicted loss-of-function variants in this cohort after filtering for artefacts caused by sequencing and annotation errors. Using an improved model of human mutation rates, we classify human protein-coding genes along a spectrum that represents tolerance to inactivation, validate this classification using data from model organisms and engineered human cells, and show that it can be used to improve the power of gene discovery for both common and rare diseases.

Modulation of prion protein expression through cryptic splice site manipulation.

Lowering expression of prion protein (PrP) is a well-validated therapeutic strategy in prion disease, but additional modalities are urgently needed. In other diseases, small molecules have proven capable of modulating pre-mRNA splicing, sometimes by forcing inclusion of cryptic exons that reduce gene expression. Here, we characterize a cryptic exon located in human PRNP's sole intron and evaluate its potential to reduce PrP expression through incorporation into the 5' untranslated region. This exon is homologous to exon 2 in nonprimate species but contains a start codon that would yield an upstream open reading frame with a stop codon prior to a splice site if included in PRNP mRNA, potentially downregulating PrP expression through translational repression or nonsense-mediated decay. We establish a minigene transfection system and test a panel of splice site alterations, identifying mutants that reduce PrP expression by as much as 78%. Our findings nominate a new therapeutic target for lowering PrP.

MRSD: A quantitative approach for assessing suitability of RNA-seq in the investigation of mis-splicing in Mendelian disease.

Variable levels of gene expression between tissues complicates the use of RNA sequencing of patient biosamples to delineate the impact of genomic variants. Here, we describe a gene- and tissue-specific metric to inform the feasibility of RNA sequencing. This overcomes limitations of using expression values alone as a metric to predict RNA-sequencing utility. We have derived a metric, minimum required sequencing depth (MRSD), that estimates the depth of sequencing required from RNA sequencing to achieve user-specified sequencing coverage of a gene, transcript, or group of genes. We applied MRSD across four human biosamples: whole blood, lymphoblastoid cell lines (LCLs), skeletal muscle, and cultured fibroblasts. MRSD has high precision (90.1%-98.2%) and overcomes transcript region-specific sequencing biases. Applying MRSD scoring to established disease gene panels shows that fibroblasts, of these four biosamples, are the optimum source of RNA for 63.1% of gene panels. Using this approach, up to 67.8% of the variants of uncertain significance in ClinVar that are predicted to impact splicing could be assayed by RNA sequencing in at least one of the biosamples. We demonstrate the utility and benefits of MRSD as a metric to inform functional assessment of splicing aberrations, in particular in the context of Mendelian genetic disorders to improve diagnostic yield.

The PS4-likelihood ratio calculator: flexible allocation of evidence weighting for case-control data in variant classification.

BACKGROUND: The 2015 American College of Medical Genetics/Association of Molecular Pathology (ACMG/AMP) variant classification framework specifies that case-control observations can be scored as 'strong' evidence (PS4) towards pathogenicity. METHODS: We developed the PS4-likelihood ratio calculator (PS4-LRCalc) for quantitative evidence assignment based on the observed variant frequencies in cases and controls. Binomial likelihoods are computed for two models, each defined by prespecified OR thresholds. Model 1 represents the hypothesis of association between variant and phenotype (eg, OR≥5) and model 2 represents the hypothesis of non-association (eg, OR≤1). RESULTS: PS4-LRCalc enables continuous quantitation of evidence for variant classification expressed as a likelihood ratio (LR), which can be log-converted into log LR (evidence points). Using PS4-LRCalc, observed data can be used to quantify evidence towards either pathogenicity or benignity. Variants can also be evaluated against models of different penetrance. The approach is applicable to balanced data sets generated for more common phenotypes and smaller data sets more typical in very rare disease variant evaluation. CONCLUSION: PS4-LRCalc enables flexible evidence quantitation on a continuous scale for observed case-control data. The converted LR is amenable to incorporation into the now widely used 2018 updated Bayesian ACMG/AMP framework.

Non-coding region variants upstream of MEF2C cause severe developmental disorder through three distinct loss-of-function mechanisms.

Clinical genetic testing of protein-coding regions identifies a likely causative variant in only around half of developmental disorder (DD) cases. The contribution of regulatory variation in non-coding regions to rare disease, including DD, remains very poorly understood. We screened 9,858 probands from the Deciphering Developmental Disorders (DDD) study for de novo mutations in the 5' untranslated regions (5' UTRs) of genes within which variants have previously been shown to cause DD through a dominant haploinsufficient mechanism. We identified four single-nucleotide variants and two copy-number variants upstream of MEF2C in a total of ten individual probands. We developed multiple bespoke and orthogonal experimental approaches to demonstrate that these variants cause DD through three distinct loss-of-function mechanisms, disrupting transcription, translation, and/or protein function. These non-coding region variants represent 23% of likely diagnoses identified in MEF2C in the DDD cohort, but these would all be missed in standard clinical genetics approaches. Nonetheless, these variants are readily detectable in exome sequence data, with 30.7% of 5' UTR bases across all genes well covered in the DDD dataset. Our analyses show that non-coding variants upstream of genes within which coding variants are known to cause DD are an important cause of severe disease and demonstrate that analyzing 5' UTRs can increase diagnostic yield. We also show how non-coding variants can help inform both the disease-causing mechanism underlying protein-coding variants and dosage tolerance of the gene.

Non-coding variants are a rare cause of recessive developmental disorders in trans with coding variants.

PURPOSE: Identifying pathogenic non-coding variants is challenging. A single protein-altering variant is often identified in a recessive gene in individuals with developmental disorders (DD), but the prevalence of pathogenic non-coding 'second hits' in trans with these is unknown. METHODS: In 4,073 genetically undiagnosed rare disease trio probands from the 100,000 Genomes project, we identified rare heterozygous protein-altering variants in recessive DD-associated genes. We identified rare non-coding variants on the other haplotype in introns, untranslated regions (UTRs), promoters, and candidate enhancer regions. We clinically evaluated the top candidates for phenotypic fit, and performed functional testing where possible. RESULTS: We identified 3,761 rare heterozygous loss-of-function or ClinVar pathogenic variants in recessive DD-associated genes in 2,430 probands. For 1,366 (36.3%) of these, we identified at least one rare non-coding variant in trans. Bioinformatic filtering and clinical review, revealed seven to be a good clinical fit. After detailed characterisation, we identified likely diagnoses for three probands (in GAA, NPHP3, and PKHD1) and candidate diagnoses in a further three (PAH, LAMA2, IGHMBP2). CONCLUSION: We developed a systematic approach to uncover new diagnoses involving compound heterozygous coding/non-coding variants and conclude that this mechanism is likely to be a rare cause of DDs.

Toward robust clinical genome interpretation: Developing a consistent terminology to characterize Mendelian disease-gene relationships-allelic requirement, inheritance modes, and disease mechanisms.

PURPOSE: The terminology used for gene-disease curation and variant annotation to describe inheritance, allelic requirement, and both sequence and functional consequences of a variant is currently not standardized. There is considerable discrepancy in the literature and across clinical variant reporting in the derivation and application of terms. Here, we standardize the terminology for the characterization of disease-gene relationships to facilitate harmonized global curation and to support variant classification within the ACMG/AMP framework. METHODS: Terminology for inheritance, allelic requirement, and both structural and functional consequences of a variant used by Gene Curation Coalition members and partner organizations was collated and reviewed. Harmonized terminology with definitions and use examples was created, reviewed, and validated. RESULTS: We present a standardized terminology to describe gene-disease relationships, and to support variant annotation. We demonstrate application of the terminology for classification of variation in the ACMG SF 2.0 genes recommended for reporting of secondary findings. Consensus terms were agreed and formalized in both Sequence Ontology (SO) and Human Phenotype Ontology (HPO) ontologies. Gene Curation Coalition member groups intend to use or map to these terms in their respective resources. CONCLUSION: The terminology standardization presented here will improve harmonization, facilitate the pooling of curation datasets across international curation efforts and, in turn, improve consistency in variant classification and genetic test interpretation.

Ethnicity, consanguinity, and genetic architecture of hypertrophic cardiomyopathy.

AIMS: Hypertrophic cardiomyopathy (HCM) is characterized by phenotypic heterogeneity that is partly explained by the diversity of genetic variants contributing to disease. Accurate interpretation of these variants constitutes a major challenge for diagnosis and implementing precision medicine, especially in understudied populations. The aim is to define the genetic architecture of HCM in North African cohorts with high consanguinity using ancestry-matched cases and controls. METHODS AND RESULTS: Prospective Egyptian patients (n = 514) and controls (n = 400) underwent clinical phenotyping and genetic testing. Rare variants in 13 validated HCM genes were classified according to standard clinical guidelines and compared with a prospective HCM cohort of majority European ancestry (n = 684). A higher prevalence of homozygous variants was observed in Egyptian patients (4.1% vs. 0.1%, P = 2 × 10-7), with variants in the minor HCM genes MYL2, MYL3, and CSRP3 more likely to present in homozygosity than the major genes, suggesting these variants are less penetrant in heterozygosity. Biallelic variants in the recessive HCM gene TRIM63 were detected in 2.1% of patients (five-fold greater than European patients), highlighting the importance of recessive inheritance in consanguineous populations. Finally, rare variants in Egyptian HCM patients were less likely to be classified as (likely) pathogenic compared with Europeans (40.8% vs. 61.6%, P = 1.6 × 10-5) due to the underrepresentation of Middle Eastern populations in current reference resources. This proportion increased to 53.3% after incorporating methods that leverage new ancestry-matched controls presented here. CONCLUSION: Studying consanguineous populations reveals novel insights with relevance to genetic testing and our understanding of the genetic architecture of HCM.

Annotating high-impact 5'untranslated region variants with the UTRannotator.

SUMMARY: Current tools to annotate the predicted effect of genetic variants are heavily biased towards protein-coding sequence. Variants outside of these regions may have a large impact on protein expression and/or structure and can lead to disease, but this effect can be challenging to predict. Consequently, these variants are poorly annotated using standard tools. We have developed a plugin to the Ensembl Variant Effect Predictor, the UTRannotator, that annotates variants in 5'untranslated regions (5'UTR) that create or disrupt upstream open reading frames. We investigate the utility of this tool using the ClinVar database, providing an annotation for 31.9% of all 5'UTR (likely) pathogenic variants, and highlighting 31 variants of uncertain significance as candidates for further follow-up. We will continue to update the UTRannotator as we gain new knowledge on the impact of variants in UTRs. AVAILABILITY AND IMPLEMENTATION: UTRannotator is freely available on Github: https://github.com/ImperialCardioGenetics/UTRannotator. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Quantifying evidence toward pathogenicity for rare phenotypes: The case of succinate dehydrogenase genes, SDHB and SDHD.

PURPOSE: The weight of the evidence to attach to observation of a novel rare missense variant in SDHB or SDHD in individuals with the rare neuroendocrine tumors, pheochromocytomas and paragangliomas (PCC/PGL), is uncertain. METHODS: We compared the frequency of SDHB and SDHD very rare missense variants (VRMVs) in 6328 and 5847 cases of PCC/PGL, respectively, with that of population controls to generate a pan-gene VRMV likelihood ratio (LR). Via windowing analysis, we measured regional enrichments of VRMVs to calculate the domain-specific VRMV-LR (DS-VRMV-LR). We also calculated subphenotypic LRs for variant pathogenicity for various clinical, histologic, and molecular features. RESULTS: We estimated the pan-gene VRMV-LR to be 76.2 (54.8-105.9) for SDHB and 14.8 (8.7-25.0) for SDHD. Clustering analysis revealed an SDHB enriched region (ɑɑ 177-260, P = .001) for which the DS-VRMV-LR was 127.2 (64.9-249.4) and an SDHD enriched region (ɑɑ 70-114, P = .000003) for which the DS-VRMV-LR was 33.9 (14.8-77.8). Subphenotypic LRs exceeded 6 for invasive disease (SDHB), head-and-neck disease (SDHD), multiple tumors (SDHD), family history of PCC/PGL, loss of SDHB staining on immunohistochemistry, and succinate-to-fumarate ratio >97 (SDHB, SDHD). CONCLUSION: Using methodology generalizable to other gene-phenotype dyads, the LRs relating to rarity and phenotypic specificity for a single observation in PCC/PGL of a SDHB/SDHD VRMV can afford substantial evidence toward pathogenicity.

Reevaluating the Genetic Contribution of Monogenic Dilated Cardiomyopathy.

BACKGROUND: Dilated cardiomyopathy (DCM) is genetically heterogeneous, with >100 purported disease genes tested in clinical laboratories. However, many genes were originally identified based on candidate-gene studies that did not adequately account for background population variation. Here we define the frequency of rare variation in 2538 patients with DCM across protein-coding regions of 56 commonly tested genes and compare this to both 912 confirmed healthy controls and a reference population of 60 706 individuals to identify clinically interpretable genes robustly associated with dominant monogenic DCM. METHODS: We used the TruSight Cardio sequencing panel to evaluate the burden of rare variants in 56 putative DCM genes in 1040 patients with DCM and 912 healthy volunteers processed with identical sequencing and bioinformatics pipelines. We further aggregated data from 1498 patients with DCM sequenced in diagnostic laboratories and the Exome Aggregation Consortium database for replication and meta-analysis. RESULTS: Truncating variants in TTN and DSP were associated with DCM in all comparisons. Variants in MYH7, LMNA, BAG3, TNNT2, TNNC1, PLN, ACTC1, NEXN, TPM1, and VCL were significantly enriched in specific patient subsets, with the last 2 genes potentially contributing primarily to early-onset forms of DCM. Overall, rare variants in these 12 genes potentially explained 17% of cases in the outpatient clinic cohort representing a broad range of adult patients with DCM and 26% of cases in the diagnostic referral cohort enriched in familial and early-onset DCM. Although the absence of a significant excess in other genes cannot preclude a limited role in disease, such genes have limited diagnostic value because novel variants will be uninterpretable and their diagnostic yield is minimal. CONCLUSIONS: In the largest sequenced DCM cohort yet described, we observe robust disease association with 12 genes, highlighting their importance in DCM and translating into high interpretability in diagnostic testing. The other genes analyzed here will need to be rigorously evaluated in ongoing curation efforts to determine their validity as Mendelian DCM genes but have limited value in diagnostic testing in DCM at present. This data will contribute to community gene curation efforts and will reduce erroneous and inconclusive findings in diagnostic testing.

Disease-specific variant pathogenicity prediction significantly improves variant interpretation in inherited cardiac conditions.

PURPOSE: Accurate discrimination of benign and pathogenic rare variation remains a priority for clinical genome interpretation. State-of-the-art machine learning variant prioritization tools are imprecise and ignore important parameters defining gene-disease relationships, e.g., distinct consequences of gain-of-function versus loss-of-function variants. We hypothesized that incorporating disease-specific information would improve tool performance. METHODS: We developed a disease-specific variant classifier, CardioBoost, that estimates the probability of pathogenicity for rare missense variants in inherited cardiomyopathies and arrhythmias. We assessed CardioBoost's ability to discriminate known pathogenic from benign variants, prioritize disease-associated variants, and stratify patient outcomes. RESULTS: CardioBoost has high global discrimination accuracy (precision recall area under the curve [AUC] 0.91 for cardiomyopathies; 0.96 for arrhythmias), outperforming existing tools (4-24% improvement). CardioBoost obtains excellent accuracy (cardiomyopathies 90.2%; arrhythmias 91.9%) for variants classified with >90% confidence, and increases the proportion of variants classified with high confidence more than twofold compared with existing tools. Variants classified as disease-causing are associated with both disease status and clinical severity, including a 21% increased risk (95% confidence interval [CI] 11-29%) of severe adverse outcomes by age 60 in patients with hypertrophic cardiomyopathy. CONCLUSIONS: A disease-specific variant classifier outperforms state-of-the-art genome-wide tools for rare missense variants in inherited cardiac conditions ( https://www.cardiodb.org/cardioboost/ ), highlighting broad opportunities for improved pathogenicity prediction through disease specificity.

Enhancing rare variant interpretation in inherited arrhythmias through quantitative analysis of consortium disease cohorts and population controls.

PURPOSE: Stringent variant interpretation guidelines can lead to high rates of variants of uncertain significance (VUS) for genetically heterogeneous disease like long QT syndrome (LQTS) and Brugada syndrome (BrS). Quantitative and disease-specific customization of American College of Medical Genetics and Genomics/Association for Molecular Pathology (ACMG/AMP) guidelines can address this false negative rate. METHODS: We compared rare variant frequencies from 1847 LQTS (KCNQ1/KCNH2/SCN5A) and 3335 BrS (SCN5A) cases from the International LQTS/BrS Genetics Consortia to population-specific gnomAD data and developed disease-specific criteria for ACMG/AMP evidence classes-rarity (PM2/BS1 rules) and case enrichment of individual (PS4) and domain-specific (PM1) variants. RESULTS: Rare SCN5A variant prevalence differed between European (20.8%) and Japanese (8.9%) BrS patients (p = 5.7 × 10-18) and diagnosis with spontaneous (28.7%) versus induced (15.8%) Brugada type 1 electrocardiogram (ECG) (p = 1.3 × 10-13). Ion channel transmembrane regions and specific N-terminus (KCNH2) and C-terminus (KCNQ1/KCNH2) domains were characterized by high enrichment of case variants and >95% probability of pathogenicity. Applying the customized rules, 17.4% of European BrS and 74.8% of European LQTS cases had (likely) pathogenic variants, compared with estimated diagnostic yields (case excess over gnomAD) of 19.2%/82.1%, reducing VUS prevalence to close to background rare variant frequency. CONCLUSION: Large case-control data sets enable quantitative implementation of ACMG/AMP guidelines and increased sensitivity for inherited arrhythmia genetic testing.