A plugin for the ensembl variant effect predictor that uses MaxEntScan to predict variant spliceogenicity

Abstract

Summary: Assessing the pathogenicity of genetic variants can be a complex and challenging task. Spliceogenic variants, which alter mRNA splicing, may yield mature transcripts that encode nonfunctional protein products, an important predictor of Mendelian disease risk. However, most variant annotation tools do not adequately assess spliceogenicity outside the native splice site and thus the disease-causing potential of variants in other intronic and exonic regions is often overlooked. Here, we present a plugin for the Ensembl Variant Effect Predictor that packages MaxEntScan and extends its functionality to provide splice site predictions using a maximum entropy model. The plugin incorporates a sliding window algorithm to predict splice site loss or gain for any variant that overlaps a transcript feature. We also demonstrate the utility of the plugin by comparing our predictions to two mRNA splicing datasets containing several cancer-susceptibility genes.