Ensembl VEP makes it easy to annotate your variants with the reference data essential to variant filtering and interpretation. This includes the results of multiplex assays of variant effect (MAVEs) which report on observed cellular phenotypes across nearly all possible variants in specific targeted regulatory or coding regions.

There are a range of different MAVE methods, all of which involve the systematic mutagenesis of a region followed by a multiplex assay which checks for impact on a specific type of function. From the assay, we can build a map scoring the effect of nearly every possible variant at every location. Many regions are being systematically assayed to build an extensive map of variant effect. When a variant allele is observed in genome sequencing, we can look up the location in the map and use the results to guide interpretation.

MaveDB accepts submissions of MAVE data which it presents in standard views and formats. Part of this work involves mapping the results of the functional experiments (which are initially reported relative to assay sequences) to accessioned reference sequences so they can be used in downstream processes. A recent paper by Arbesfeld et al describes this process and how results were integrated into resources including Ensembl.

(a) Ensembl VEP results with information from MaveDB including links to the publication and links to the assay page on the MaveDB website; (b) a MaveDB plot of the tolerance to change across the PTEN gene and (c) an effect map for PTEN from MaveDB showing the individual variant impact over the first 48 amino acids.

​​Ensembl VEP currently reports results from over 1000 regional assays. As methods vary, scores are not comparable across different regions with different assays. We recommend reading the publication, which can be accessed via a link from the Ensembl VEP results table, to further understand the assay’s scoring range and relevance to your analysis. When using MAVE data to investigate the clinical impact of a variant, it’s important to ensure the assay measures a function related to the disease mechanism. Resources like Gene2Phenotype curate and share information on gene-disease mechanisms. For guidelines on using functional data in clinical variant interpretation, see Brnich et al. 

To access MaveDB data via the Ensembl VEP web interface, select the ‘MaveDB’ option in the ‘Functional Effect’ section. The results table will include any scores available for the variants and links to MaveDB (Figure a) where you can find heatmaps of variant effect and descriptions of how the assay was performed, which can be used to pick assay-specific score cutoffs (Figure b and c). Use the option mavedb=1 with the REST API to see results. The MaveDB plugin can be used for command line analysis with data downloaded from https://ftp.ensembl.org/pub/current_variation/MaveDB/.

Authors: Sarah Hunt and Jamie Allen