The Ensembl Variant Effect Predictor

BackgroundVariant annotation is a crucial step in the analysis of genome sequencing data. Functional annotation results can have a strong influence on the ultimate conclusions of disease studies. Incorrect or incomplete annotations can cause researchers both to overlook potentially disease-relevant DNA variants and to dilute interesting variants in a pool of false positives. Researchers are aware of these issues in general, but the extent of the dependency of final results on the choice of transcripts and software used for annotation has not been quantified in detail.MethodsThis paper quantifies the extent of differences in annotation of 80 million variants from a whole-genome sequencing study. We compare results using the RefSeq and Ensembl transcript sets as the basis for variant annotation with the software Annovar, and also compare the results from two annotation software packages, Annovar and VEP (Ensembl’s Variant Effect Predictor), when using Ensembl transcripts.ResultsWe found only 44% agreement in annotations for putative loss-of-function variants when using the RefSeq and Ensembl transcript sets as the basis for annotation with Annovar. The rate of matching annotations for loss-of-function and nonsynonymous variants combined was 79% and for all exonic variants it was 83%. When comparing results from Annovar and VEP using Ensembl transcripts, matching annotations were seen for only 65% of loss-of-function variants and 87% of all exonic variants, with splicing variants revealed as the category with the greatest discrepancy. Using these comparisons, we characterised the types of apparent errors made by Annovar and VEP and discuss their impact on the analysis of DNA variants in genome sequencing studies.ConclusionsVariant annotation is not yet a solved problem. Choice of transcript set can have a large effect on the ultimate variant annotations obtained in a whole-genome sequencing study. Choice of annotation software can also have a substantial effect. The annotation step in the analysis of a genome sequencing study must therefore be considered carefully, and a conscious choice made as to which transcript set and software are used for annotation.

Genetic diseases are mostly implicated with genetic variants, including missense, synonymous, non-sense, and copy number variants. These different kinds of variants are indicated to affect phenotypes in various ways from previous studies. It remains essential but challenging to understand the functional consequences of these genetic variants, especially the noncoding ones, due to the lack of corresponding annotations. While many computational methods have been proposed to identify the risk variants. Most of them have only curated DNA-level and protein-level annotations to predict the pathogenicity of the variants, and others have been restricted to missense variants exclusively. In this study, we have curated DNA-, RNA-, and protein-level features to discriminate disease-causing variants in both coding and noncoding regions, where the features of protein sequences and protein structures have been shown essential for analyzing missense variants in coding regions while the features related to RNA-splicing and RBP binding are significant for variants in noncoding regions and synonymous variants in coding regions. Through the integration of these features, we have formulated the Multi-level feature Genomic Variants Predictor (ML-GVP) using the gradient boosting tree. The method has been trained on more than 400,000 variants in the Sherloc-training set from the 6th critical assessment of genome interpretation with superior performance. The method is one of the two best-performing predictors on the blind test in the Sherloc assessment, and is further confirmed by another independent test dataset of de novo variants.

The Ensembl Variant Effect Predictor (VEP) is a freely available, open-source tool for the annotation and filtering of genomic variants. It predicts variant molecular consequences using the Ensembl/GENCODE or RefSeq gene sets. It also reports phenotype associations from databases such as ClinVar, allele frequencies from studies including gnomAD, and predictions of deleteriousness from tools such as Sorting Intolerant From Tolerant and Combined Annotation Dependent Depletion. Ensembl VEP includes filtering options to customize variant prioritization. It is well supported and updated roughly quarterly to incorporate the latest gene, variant, and phenotype association information. Ensembl VEP analysis can be performed using a highly configurable, extensible command-line tool, a Representational State Transfer application programming interface, and a user-friendly web interface. These access methods are designed to suit different levels of bioinformatics experience and meet different needs in terms of data size, visualization, and flexibility. In this tutorial, we will describe performing variant annotation using the Ensembl VEP web tool, which enables sophisticated analysis through a simple interface.

Accurate clinical interpretation of genetic variants requires integration of functional predictions, evolutionary constraint, population allele frequencies, and clinical evidence from heterogeneous resources. Conventional workflows based on standalone tools such as ensembl variant effect predictor (VEP) and ANNOVAR require complex manual configuration of plugins and databases, generate verbose transcript-level outputs unsuitable for clinical review, and rely on ad hoc scripts for format conversion and prioritization. These limitations hinder reproducibility and scalability, making data interpretation a major bottleneck in genomic medicine. We present MuSA (Multi-Source variant Annotation), an nf-core-compliant Nextflow pipeline that automates germline variant annotation from resource setup to clinical interpretation. MuSA supports both a streamlined basic mode for diagnostic workflows and an extended deep-annotation mode for comprehensive analyses. The pipeline integrates Ensembl VEP with 22 curated plugins (including AlphaMissense, CADD, SpliceAI, and Enformer), ANNOVAR, a standalone pre-configured dbNSFP distribution, the RENOVO pathogenicity predictor, and automated ACMG/AMP classification via GeneBe and InterVar. MuSA standardizes input VCFs, executes parallel annotation branches in a fully containerized workflow, and consolidates results into richly annotated mutation annotation format (MAF) files (up to 920 columns per variant), alongside interactive HTML reports tailored for clinical review with HPO-matched gene panels. Benchmarked on a WES-like dataset of 22,705 variants derived from the public GIAB NA12878/HG001 GRCh38 benchmark VCF, MuSA completes full extended-mode annotation in approximately 20min on a 64-core server. Systematic comparison with nf-core/sarek and nf-core/variantprioritization demonstrates that MuSA uniquely combines automated resource management with YAML-based version tracking and SHA-256 integrity verification, native dbNSFP integration, RENOVO-based VUS prioritization, HPO-driven gene panel filtering, and a clinically oriented interactive HTML report; those features are mostly absent in existing nf-core annotation pipelines. Containerization through Docker/Singularity and predefined execution profiles support reproducible deployment across workstations, HPC clusters, and cloud environments. MuSA provides an end-to-end framework for clinically oriented germline variant annotation and prioritization, addressing key limitations of manual and general-purpose workflows. Its dual-output design bridges research (machine-readable MAF files compatible with downstream tools such as maftools) and clinical diagnostics (interpretation-ready HTML reports), supporting reproducible and standardized variant interpretation across teams. Current limitations include restriction to germline small variants on hg38, a substantial storage footprint (up to 223.5 GB for extended mode), and dependence on external APIs for ACMG/AMP classification and phenotype-driven filtering. The RENOVO-based VUS prioritization module is experimental and requires expert review before clinical interpretation.

High-throughput DNA sequencing has become a mainstay for the discovery of genomic variants that may cause disease or affect phenotype. A next-generation sequencing pipeline typically identifies thousands of variants in each sample. A particular challenge is the annotation of each variant in a way that is useful to downstream consumers of the data, such as clinical sequencing centers or researchers. These users may require that all data storage and analysis remain on secure local servers to protect patient confidentiality or intellectual property, may have unique and changing needs to draw on a variety of annotation data sets and may prefer not to rely on closed-source applications beyond their control. Here we describe scalable methods for using the plugin capability of the Ensembl Variant Effect Predictor to enrich its basic set of variant annotations with additional data on genes, function, conservation, expression, diseases, pathways and protein structure, and describe an extensible framework for easily adding additional custom data sets.

In cancer genomics, precise variant annotation is crucial for clinical decisions, drug development, and research. The burgeoning genomic data offers an opportunity to use data-driven approaches to generate knowledge that supports clinical decisions. Particularly machine learning (ML) and Deep learning (DL), are becoming essential, as their application is fast, scalable, simple to implement, and generates reproducible results. The methods ranging from simple sequence-based alignment scoring to advanced algorithms like Logistic regression, Support vector machine, and Recurrent neural networks, have been employed by multiple variant annotation tools. This study compares in-silico methods available in Ensembl's Variant Effect Predictor (VEP) using a test dataset with COSMIC annotations. ML/DL success relies on robust training sets with comprehensive genomic variants features, including effects on transcription/translation, genomic context, annotation resources, in silico pathogenicity predictions, and population allele frequency. The training set, composed of known benign or pathogenic variants, serves as a reference for these algorithms to classify new and unseen variants. Our analysis reveals a limited concordance between the prediction algorithms. Despite comparable true/false positives/negatives, discrepancies persist in variant classification. Certain algorithms exhibit a propensity to over or under-call deleterious mutations. Some demonstrate a tendency to classify random variants in non-cancer genes as deleterious. Challenges include the absence of consensus on informative features, diverse training datasets, and restriction to well annotated proteins/transcripts. Balancing the sensitivity and false positives in detecting cancer drivers is crucial. Integrating individual prediction scores with ML algorithms enhances tool performance but comes with risk of error propagation, and limited accuracy. The study emphasises the need for context-specific variant classification tools, as many variants' impacts are cancer-type specific, and some may drive disease synergistically. Existing tools, designed for a "one variant - one score approach," struggle to capture complex associations, especially those dependent on changes in the tumour microenvironment. Highlighting areas for improvement, the study addresses the "black box" problem in decision processes. While limited interpretability might not hinder practical applications, tools should evolve to assess more complex associations guided by biology. Formal consensus, reference training datasets, and standards are deemed essential for developing next-generation tools. The envisioned context-dependent tools aim to streamline feature complexity, thereby mitigating the black box problem and advancing the accuracy and interpretability of cancer variant annotation. Citation Format: Madhumita, Zbyslaw Sondka, Jon Teague. Evaluating the utility of in silico variant annotation tools for cancer driver detection [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4884.

Cystic Fibrosis Transmembrane Conductance Regulator Gene Variations in Coding and Noncoding Regions in Congenital Bilateral Absence of the Vas Deferens Dependent Infertility

A performance evaluation study: Variant annotation tools - the enigma of clinical next generation sequencing (NGS) based genetic testing

The interpretation of variants in cancer is frequently focused on direct protein coding alterations. However, most somatic mutations are in noncoding regions of the genome, and even exonic mutations may have unidentified noncoding consequences. Here we present Regtools, a software package designed to efficiently identify variants that may cause aberrant splicing in tumors. Our tool integrates variant calls from genomic data with junctions extracted from transcriptomic data in order to examine potential cis alterations to splicing near a somatic variant. Based on user-defined parameters and position relative to known exons, variants are first annotated as splicing relevant or not. Splicing junctions are inferred from transcriptomic sequencing data, and comparison of junctions to a reference transcriptome allows for identification and annotation of novel junctions and nearby regulatory or splicing motifs. From there, mutations are associated with junctions that overlap with a flanking region. In order to evaluate Regtools, we used it to analyze the transcriptional output of tumor-sequencing data from two cohorts of cancer patients, one of hepatocellular carcinoma and one of small cell lung cancer with 28 and 21 samples, respectively. We performed whole-exome and RNA sequencing on each sample. Somatic variants were called on whole-exome alignment data. For each cohort, we compared the junctional profiles between tumors and identified numerous examples of variants for which there are elevated levels of proximal novel or known junctions. Moreover, out of 754 (153 in HCC; 601 in SCLC) variants identified as splicing relevant by our approach, only 165 (20 in HCC; 145 in SCLC) were annotated as splicing relevant by Ensembl's Variant Effect Predictor, using the recommended “per_gene” option. This preliminary analysis illustrates the importance of an efficient, user-friendly computational tool for identifying important noncoding variants that would otherwise be undervalued or perhaps even completely ignored by traditional methods and annotators. Regtools is freely available and open source (https://github.com/griffithlab/regtools). Citation Format: Yang-Yang Feng, Avinash Ramu, Zachary L. Skidmore, Jason Kunisaki, Kelsy C. Cotto, Obi L. Griffith, Malachi Griffith. Regtools: Integrated analysis of genomic and transcriptomic data for discovery of mutations associated with aberrant splicing in cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2285.

BackgroundSevere equine asthma is a chronic inflammatory disease of the lung in horses similar to low-Th2 late-onset asthma in humans. This study aimed to determine the utility of RNA-Seq to call gene sequence variants, and to identify sequence variants of potential relevance to the pathogenesis of asthma.MethodsRNA-Seq data were generated from endobronchial biopsies collected from six asthmatic and seven non-asthmatic horses before and after challenge (26 samples total). Sequences were aligned to the equine genome with Spliced Transcripts Alignment to Reference software. Read preparation for sequence variant calling was performed with Picard tools and Genome Analysis Toolkit (GATK). Sequence variants were called and filtered using GATK and Ensembl Variant Effect Predictor (VEP) tools, and two RNA-Seq predicted sequence variants were investigated with both PCR and Sanger sequencing. Supplementary analysis of novel sequence variant selection with VEP was based on a score of <0.01 predicted with Sorting Intolerant from Tolerant software, missense nature, location within the protein coding sequence and presence in all asthmatic individuals. For select variants, effect on protein function was assessed with Polymorphism Phenotyping 2 and screening for non-acceptable polymorphism 2 software. Sequences were aligned and 3D protein structures predicted with Geneious software. Difference in allele frequency between the groups was assessed using a Pearson’s Chi-squared test with Yates’ continuity correction, and difference in genotype frequency was calculated using the Fisher’s exact test for count data.ResultsRNA-Seq variant calling and filtering correctly identified substitution variants in PACRG and RTTN. Sanger sequencing confirmed that the PACRG substitution was appropriately identified in all 26 samples while the RTTN substitution was identified correctly in 24 of 26 samples. These variants of uncertain significance had substitutions that were predicted to result in loss of function and to be non-neutral. Amino acid substitutions projected no change of hydrophobicity and isoelectric point in PACRG, and a change in both for RTTN. For PACRG, no difference in allele frequency between the two groups was detected but a higher proportion of asthmatic horses had the altered RTTN allele compared to non-asthmatic animals.DiscussionRNA-Seq was sensitive and specific for calling gene sequence variants in this disease model. Even moderate coverage (<10–20 counts per million) yielded correct identification in 92% of samples, suggesting RNA-Seq may be suitable to detect sequence variants in low coverage samples. The impact of amino acid alterations in PACRG and RTTN proteins, and possible association of the sequence variants with asthma, is of uncertain significance, but their role in ciliary function may be of future interest.

Background. Severe equine asthma is a chronic inflammatory disease of the lung in horses similar to low-Th2 late-onset asthma in humans. This study aimed to determine the utility of RNA-Seq to call gene sequence variants, and to identify sequence variants or potential relevance to the pathogenesis of asthma. Methods. RNA-Seq data were generated from endobronchial biopsies collected from 6 asthmatic and 7 non-asthmatic horses before and after challenge (26 samples total). Sequences were aligned to the equine genome with Spliced Transcripts Alignment to Reference software. Read preparation for sequence variant calling was performed with Picard tools and Genome Analysis Toolkit (GATK). Sequence variants were called and filtered using GATK and Ensembl Variant Effect Predictor (VEP) tools, and two RNA-Seq predicted sequence variants were investigated with both PCR and Sanger sequencing. Supplementary analysis of novel sequence variant selection with VEP was based on a score of &lt;0.01 predicted with Sorting Intolerant From Tolerant (SIFT) software, missense nature, location within the protein coding sequence and presence in all asthmatic individuals. For select variants, effect on protein function was assessed with Polymorphism Phenotyping (PolyPhen) 2 and Screening for Non-Acceptable Polymorphism (SNAP) 2 software. Sequences were aligned and 3D protein structures predicted with Geneious software. Difference in allele frequency between the groups was assessed using a Pearson's Chi-squared test with Yates' continuity correction, and difference in genotype frequency was calculated using the Fisher's exact test for count data. Results. RNA-Seq variant calling and filtering correctly identified substitution variants in PACRG and RTTN. Sanger sequencing confirmed that the PACRG substitution was appropriately identified in all 26 samples while the RTTN substitution was identified correctly in 24 of 26 samples. These variants of uncertain significance had substitutions that were predicted to result in loss of function and to be non-neutral. Amino acid substitutions projected no change of hydrophobicity and isoelectric point in PACRG, and a change in both for RTTN. For PACRG, no difference in allele frequency between the two groups was detected but a higher proportion of asthmatic horses had the altered RTTN allele compared to non-asthmatic animals. Discussion. RNA-Seq was sensitive and specific for calling gene sequence variants in this disease model. Even moderate coverage (&lt;10-20 cpm) yielded correct identification in 92% of samples, suggesting RNA-Seq may be suitable to detect sequence variants in low coverage samples. The impact of amino acid alterations in PACRG and RTTN proteins, and possible association of the sequence variants with asthma, is of uncertain significance, but their role in ciliary function may be of future interest.

Utilizing data from published tuberculosis (TB) genome-wide association studies (GWAS), we use a bioinformatics pipeline to detect all polymorphisms in linkage disequilibrium (LD) with variants previously implicated in TB disease susceptibility. The probability that these variants had a predicted regulatory function was estimated using RegulomeDB and Ensembl’s Variant Effect Predictor. Subsequent genotyping of these 133 predicted regulatory polymorphisms was performed in 400 admixed South African TB cases and 366 healthy controls in a population-based case-control association study to fine-map the causal variant. We detected associations between tuberculosis susceptibility and six intronic polymorphisms located in MARCO, IFNGR2, ASHAS2, ACACA, NISCH and TLR10. Our post-GWAS approach demonstrates the feasibility of combining multiple TB GWAS datasets with linkage information to identify regulatory variants associated with this infectious disease.

Annotation of human genetic variants enables genotype-phenotype association studies at the gene, pathway, and tissue level. Annotation results are difficult to reproduce across study sites due to shifting software versions and a lack of a unified hardware interface between study sites. Cloud computing offers a promising solution by integrating hardware and software into reproducible virtual appliances which may be utilized on-demand and shared across institutions. We developed ENSEMBL VEP on EC2 (EVE), a cloud-based virtual appliance for annotation of human genetic variants built around the ENSEMBL Variant Effect Predictor. We integrated virtual hardware infrastructure, open-source software, and publicly available genomic datasets to provide annotation capability for genetic variants in the context of genes/transcripts, Gene Ontology pathways, tissue-specific expression from the Gene Expression Atlas, miRNA annotations, minor allele frequencies from the 1000 Genomes Project and the Exome Aggregation Consortium, and deleteriousness scores from Combined Annotation Dependent Depletion. We demonstrate the utility of EVE by annotating the genetic variants in a case-control study of glaucoma. Cloud computing can reduce the difficulty of replicating complex software pipelines such as annotation pipelines across study sites. We provide a publicly available CloudFormation template of the EVE virtual appliance which can automatically provision and deploy a parameterized, preconfigured hardware/software stack ready for annotation of human genetic variants (github.com/epistasislab/EVE). This approach offers increased reproducibility in human genetic studies by providing a unified appliance to researchers across the world.

Non-syndromic cleft lip with or without cleft palate (NSCL/P) is one of the most common congenital craniofacial malformations, characterized by significant genetic heterogeneity and polygenic features. Historically, genetic studies of NSCL/P have primarily focused on coding region variants. However, only approximately 2% of the human genome consists of coding sequences, with the majority of variants located in non-coding regions. Recent large-scale genome-wide association studies have indicated that risk loci associated with NSCL/P are predominantly enriched in regulatory elements within non-coding regions of the genome. These non-coding variants, while not directly altering protein structures, can influence gene expression patterns in specific tissues or developmental spatiotemporal contexts by affecting the function of regulatory elements, thereby contributing to disease phenotypes. With the continuous advancement of functional genomics and bioinformatics methodologies, the identification, functional annotation, and mechanistic analysis of regulatory variants in non-coding regions have become critical directions in NSCL/P genetic research. This review systematically summarizes the research progress in identification methods, functional annotation strategies, and functional validation techniques for non-coding variants in NSCL/P patients. It elaborates on the significant role of non-coding regulatory elements in the pathogenesis of NSCL/P, aiming to provide new perspectives for understanding the functional significance of non-coding regions in NSCL/P and other complex diseases.

BackgroundGenomic variant prioritisation is one of the most significant bottlenecks to mainstream genomic testing in healthcare. Tools to improve precision while ensuring high recall are critical to successful mainstream clinical...