Inheritance-mode specific pathogenicity prioritization (ISPP) for human protein coding genes

Genome-wide prediction of dominant and recessive neurodevelopmental disorder-associated genes

Despite great progress, thousands of neurodevelopmental disorder (NDD) risk genes remain to be discovered. We present a computational approach that accelerates NDD risk gene identification using machine learning. First, we demonstrate that models trained solely on single-cell RNA sequencing data can robustly predict genes implicated in autism spectrum disorder (ASD), developmental and epileptic encephalopathy (DEE), and developmental delay (DD). Notably, we find differences in gene expression patterns of genes with monoallelic and bi-allelic inheritance patterns in the developing human cortex. We then integrate expression data with 300 orthogonal features, including intolerance metrics, protein-protein interaction data, and others, in a semi-supervised machine learning framework (mantis-ml) to train inheritance-specific models for these disorders. The models have high predictive power (area under the receiver operator curves [AUCs]: 0.84-0.95), and the top-ranked genes were up to 2-fold (monoallelic models) and 6-fold (bi-allelic models) more enriched for high-confidence NDD risk genes compared to genic intolerance metrics alone. Additionally, genes ranking in the top decile were 45 to 180 times more likely to have literature support than those in the bottom decile. Collectively, this work provides robust NDD risk gene predictions that can complement large-scale gene discovery efforts and underscores the importance of considering inheritance in gene risk prediction.

Ensemble and consensus approaches to prediction of recessive inheritance for missense variants in human disease

Mode of inheritance (MOI) is necessary for clinical interpretation of pathogenic variants; however, the majority of variants lack this information. Furthermore, variant effect predictors are fundamentally insensitive to recessive-acting diseases. Here, we present MOI-Pred, a variant pathogenicity prediction tool that accounts for MOI, and ConMOI, a consensus method that integrates variant MOI predictions from three independent tools. MOI-Pred integrates evolutionary and functional annotations to produce variant-level predictions that are sensitive to both dominant-acting and recessive-acting pathogenic variants. Both MOI-Pred and ConMOI show state-of-the-art performance on standard benchmarks. Importantly, dominant and recessive predictions from both tools are enriched in individuals with pathogenic variants for dominant- and recessive-acting diseases, respectively, in a real-world electronic health record (EHR)-based validation approach of 29,981 individuals. ConMOI outperforms its component methods in benchmarking and validation, demonstrating the value of consensus among multiple prediction methods. Predictions for all possible missense variants are provided in the "Data and code availability" section.

Faster and more accurate pathogenic combination predictions with VarCoPP2.0

BACKGROUND

The prediction of potentially pathogenic variant combinations in patients remains a key task in the field of medical genetics for the understanding and detection of oligogenic/multilocus diseases. Models tailored towards such cases can help shorten the gap of missing diagnoses and can aid researchers in dealing with the high complexity of the derived data. The predictor VarCoPP (Variant Combinations Pathogenicity Predictor) that was published in 2019 and identified potentially pathogenic variant combinations in gene pairs (bilocus variant combinations), was the first important step in this direction. Despite its usefulness and applicability, several issues still remained that hindered a better performance, such as its False Positive (FP) rate, the quality of its training set and its complex architecture.

RESULTS

We present VarCoPP2.0: the successor of VarCoPP that is a simplified, faster and more accurate predictive model identifying potentially pathogenic bilocus variant combinations. Results from cross-validation and on independent data sets reveal that VarCoPP2.0 has improved in terms of both sensitivity (95% in cross-validation and 98% during testing) and specificity (5% FP rate). At the same time, its running time shows a significant 150-fold decrease due to the selection of a simpler Balanced Random Forest model. Its positive training set now consists of variant combinations that are more confidently linked with evidence of pathogenicity, based on the confidence scores present in OLIDA, the Oligogenic Diseases Database ( https://olida.ibsquare.be ). The improvement of its performance is also attributed to a more careful selection of up-to-date features identified via an original wrapper method. We show that the combination of different variant and gene pair features together is important for predictions, highlighting the usefulness of integrating biological information at different levels.

CONCLUSIONS

Through its improved performance and faster execution time, VarCoPP2.0 enables a more accurate analysis of larger data sets linked to oligogenic diseases. Users can access the ORVAL platform ( https://orval.ibsquare.be ) to apply VarCoPP2.0 on their data.

An overview of germline variations in genes of primary immunodeficiences through integrative analysis of ClinVar, HGMD® and dbSNP databases

Primary immunodeficiencies (PID) are a diverse group of genetic disorders caused by inadequate development and function of immune system. Identifying genetic etiology is important for genetic counselling and treatment decisions. Clinical relevance of genetic variants is a complex problem depending on gene-specific and variant specific genotype-phenotype interactions. To address this challenge, we aimed to characterize the pathogenic landscape of PID genes by combining the analysis of germline variations reported in ClinVar and HGMD^® and identification of damaging variations available in dbSNP. We generated a joint ClinVar/HGMD database, which included 111,940 variants, among them 32,452 were classified as pathogenic/likely pathogenic. From a total of 5,415,794 bi- or multiallelic variants in PID genes recorded in dbSNP, we retrieved 38,291 high impact (HI) biallelic variants with presumably disruptive impact in the protein, of them 25,500 variants were not present in ClinVar/HGMD. Using a functional prediction algorithm, we additionally identified 28,507 deleterious and 56,016 neutral missense variants among dbSNP variants and created a collection of damaging and neutral variations in PID genes, not currently present in ClinVar/HGMD, with their allele frequencies and mappings to protein domains. The distribution of pathogenic variants from ClinVar/HGMD, HI variants and deleterious missense variants from dbSNP was analyzed in the context of hereditary pattern and gene specific metrics, such as pLI and haploinsufficiency. Our report summarized data on complex gene-specific variability in PID genes and might be useful for the identification of the most promising variants and gene regions for further study.

Essentiality-specific pathogenicity prioritization gene score to improve filtering of disease sequence data

The causal genetic variants underlying more than 50% of single gene (monogenic) disorders are yet to be discovered. Many patients with conditions likely to have a monogenic basis do not receive a confirmed molecular diagnosis which has potential impacts on clinical management. We have developed a gene-specific score, essentiality-specific pathogenicity prioritization (ESPP), to guide the recognition of genes likely to underlie monogenic disease variation to assist in filtering of genome sequence data. When a patient genome is sequenced, there are frequently several plausibly pathogenic variants identified in different genes. Recognition of the single gene most likely to include pathogenic variation can guide the identification of a causal variant. The ESPP score integrates gene-level scores which are broadly related to gene essentiality. Previous work towards the recognition of monogenic disease genes proposed a model with increasing gene essentiality from 'non-essential' to 'essential' genes (for which pathogenic variation may be incompatible with survival) with genes liable to contain disease variation positioned between these two extremes. We demonstrate that the ESPP score is useful for recognizing genes with high potential for pathogenic disease-related variation. Genes classed as essential have particularly high scores, as do genes recently recognized as strong candidates for developmental disorders. Through the integration of individual gene-specific scores, which have different properties and assumptions, we demonstrate the utility of an essentiality-based gene score to improve sequence genome filtering.

Cancer gene mutations in congenital pulmonary airway malformation patients

Background

Newborns affected with congenital pulmonary airway malformations (CPAMs) may present with severe respiratory distress or remain asymptomatic. While surgical resection is the definitive treatment for symptomatic CPAMs, prophylactic elective surgery may be recommended for asymptomatic CPAMs owing to the risk of tumour development. However, the implementation of prophylactic surgery is quite controversial on the grounds that more evidence linking CPAMs and cancer is needed. The large gap in knowledge of CPAM pathogenesis results in uncertainties and controversies in disease management. As developmental genes control postnatal cell growth and contribute to cancer development, we hypothesised that CPAMs may be underlain by germline mutations in genes governing airways development.

Methods

Sequencing of the exome of 19 patients and their unaffected parents.

Results

A more than expected number of mutations in cancer genes (false discovery rate q-value <5.01×10⁻⁵) was observed. The co-occurrence, in the same patient, of damaging variants in genes encoding interacting proteins is intriguing, the most striking being thyroglobulin (TG) and its receptor, megalin (LRP2). Both genes are highly relevant in lung development and cancer.

Conclusions

The overall excess of mutations in cancer genes may account for the reported association of CPAMs with carcinomas and provide some evidence to argue for prophylactic surgery by some surgeons.

Congenital pulmonary airway malformation (CPAM) patients have more than expected numbers of damaging variants in genes involved in lung carcinoma; this may provide evidence for clinicians choosing to adopt prophylactic excision in CPAMhttp://ow.ly/h1AE30n4DIe

De novo mutations in Caudal Type Homeo Box transcription Factor 2 (CDX2) in patients with persistent cloaca

The cloaca is an embryonic cavity that is divided into the urogenital sinus and rectum upon differentiation of the cloacal epithelium triggered by tissue-specific transcription factors including CDX2. Defective differentiation leads to persistent cloaca in humans (PC), a phenotype recapitulated in Cdx2 mutant mice. PC is linked to hypo/hyper-vitaminosis A. Although no gene has ever been identified, there is a strong evidence for a genetic contribution to PC. We applied whole-exome sequencing and copy-number-variants analyses to 21 PC patients and their unaffected parents. The damaging p.Cys132* and p.Arg237His de novo CDX2 variants were identified in two patients. These variants altered the expression of CYP26A1, a direct CDX2 target encoding the major retinoic acid (RA)-degrading enzyme. Other RA genes, including the RA-receptor alpha, were also mutated. Genes governing the development of cloaca-derived structures were recurrently mutated and over-represented in the basement-membrane components set (q-value < 1.65 × 10-6). Joint analysis of the patients' profile highlighted the extracellular matrix-receptor interaction pathway (MsigDBID: M7098, FDR: q-value < 7.16 × 10-9). This is the first evidence that PC is genetic, with genes involved in the RA metabolism at the lead. Given the CDX2 de novo variants and the role of RA, our observations could potentiate preventive measures. For the first time, a gene recapitulating PC in mouse models is found mutated in humans.

Robust and rapid algorithms facilitate large-scale whole genome sequencing downstream analysis in an integrative framework

Whole genome sequencing (WGS) is a promising strategy to unravel variants or genes responsible for human diseases and traits. However, there is a lack of robust platforms for a comprehensive downstream analysis. In the present study, we first proposed three novel algorithms, sequence gap-filled gene feature annotation, bit-block encoded genotypes and sectional fast access to text lines to address three fundamental problems. The three algorithms then formed the infrastructure of a robust parallel computing framework, KGGSeq, for integrating downstream analysis functions for whole genome sequencing data. KGGSeq has been equipped with a comprehensive set of analysis functions for quality control, filtration, annotation, pathogenic prediction and statistical tests. In the tests with whole genome sequencing data from 1000 Genomes Project, KGGSeq annotated several thousand more reliable non-synonymous variants than other widely used tools (e.g. ANNOVAR and SNPEff). It took only around half an hour on a small server with 10 CPUs to access genotypes of ∼60 million variants of 2504 subjects, while a popular alternative tool required around one day. KGGSeq's bit-block genotype format used 1.5% or less space to flexibly represent phased or unphased genotypes with multiple alleles and achieved a speed of over 1000 times faster to calculate genotypic correlation.

Burden of rare variants in ALS genes influences survival in familial and sporadic ALS

Genetic variants are implicated in the development of amyotrophic lateral sclerosis (ALS), but it is unclear whether the burden of rare variants in ALS genes has an effect on survival. We performed whole genome sequencing on 8 familial ALS (FALS) patients with superoxide dismutase 1 (SOD1) mutation and whole exome sequencing on 46 sporadic ALS (SALS) patients living in Hong Kong and found that 67% had at least 1 rare variant in the exons of 40 ALS genes; 22% had 2 or more. Patients with 2 or more rare variants had lower probability of survival than patients with 0 or 1 variant (p = 0.001). After adjusting for other factors, each additional rare variant increased the risk of respiratory failure or death by 60% (p = 0.0098). The presence of the rare variant was associated with the risk of ALS (Odds ratio 1.91, 95% confidence interval 1.03-3.61, p = 0.03), and ALS patients had higher rare variant burden than controls (MB, p = 0.004). Our findings support an oligogenic basis with the burden of rare variants affecting the development and survival of ALS.

cepip: context-dependent epigenomic weighting for prioritization of regulatory variants and disease-associated genes

It remains challenging to predict regulatory variants in particular tissues or cell types due to highly context-specific gene regulation. By connecting large-scale epigenomic profiles to expression quantitative trait loci (eQTLs) in a wide range of human tissues/cell types, we identify critical chromatin features that predict variant regulatory potential. We present cepip, a joint likelihood framework, for estimating a variant’s regulatory probability in a context-dependent manner. Our method exhibits significant GWAS signal enrichment and is superior to existing cell type-specific methods. Furthermore, using phenotypically relevant epigenomes to weight the GWAS single-nucleotide polymorphisms, we improve the statistical power of the gene-based association test.