Better prediction of functional effects for sequence variants

Abstract

Elucidating the effects of naturally occurring genetic variation is one of the major challenges for personalized health and personalized medicine. Here, we introduce SNAP2, a novel neural network based classifier that improves over the state-of-the-art in distinguishing between effect and neutral variants. Our method's improved performance results from screening many potentially relevant protein features and from refining our development data sets. Cross-validated on >100k experimentally annotated variants, SNAP2 significantly outperformed other methods, attaining a two-state accuracy (effect/neutral) of 83%. SNAP2 also outperformed combinations of other methods. Performance increased for human variants but much more so for other organisms. Our method's carefully calibrated reliability index informs selection of variants for experimental follow up, with the most strongly predicted half of all effect variants predicted at over 96% accuracy. As expected, the evolutionary information from automatically generated multiple sequence alignments gave the strongest signal for the prediction. However, we also optimized our new method to perform surprisingly well even without alignments. This feature reduces prediction runtime by over two orders of magnitude, enables cross-genome comparisons, and renders our new method as the best solution for the 10-20% of sequence orphans. SNAP2 is available at: https://rostlab.org/services/snap2webDefinitions usedDelta, input feature that results from computing the difference feature scores for native amino acid and feature scores for variant amino acid; nsSNP, non-synoymous SNP; PMD, Protein Mutant Database; SNAP, Screening for non-acceptable polymorphisms; SNP, single nucleotide polymorphism; variant, any amino acid changing sequence variant.

Partial growth hormone insensitivity and dysregulatory immune disease associated with de novo germline activating STAT3 mutations

“…For this purpose, different bioinformatics tools were used: PolyPhen-2 (Adzhubei et al, 2010), SIFT (Kumar et al, 2009), Mutation Taster (Schwarz et al, 2014), MutPred (Li et al, 2009), and SNAP 2 (Hecht et al, 2015). The following NCBI reference sequences were used: NG_007370.1 (gene), NM_139,276.2 (mRNA) and NP_40,763.1 (protein).…”

Whole Genome Sequencing and Analysis of Godawee, a Salt Tolerant Indica Rice Variety

“…SIFT (http://sift.jcvi.org/) [34], Provean (http://provean.jcvi.org) [35] and SNAP (https://rostlab.org/services/snap/) [36] algorithms were used to determine the functional effect of non-synonymous SNPs.…”

Tools for Predicting the Functional Impact of Nonsynonymous Genetic Variation

“…In general, many different machinelearning models perform equally well; for example, the latest version of SNAP averages .10 different predictors with similar performance, for which evolutionary conservation is the only feature in common to all 10 (Hecht et al 2015). The advantage of machine-learning approaches is that they can include features of very different types, and potentially a large number of such features that can be combined in highly complex and nonlinear ways.…”

“…For example, for SNPs&GO (Capriotti et al 2013b) the same learning procedure was followed with and without the additional Gene Ontology information (Ashburner et al 2000) to quantitate the improvement in prediction accuracy from this one additional data source. SNAP2 (Hecht et al 2015) reported results for multiple different machine-learning methods using the same set of features. The FATHMM article (Shihab et al 2013) reported the performance of their method before and after incorporating protein domain-specific pathogenicity weights.…”