PubMed Highlight: Evaluation of bioinformatic tools for prediction of functional impact of missense variants

This interesting paper evaluates the performance (sensitivity and specificity) of 9 different tools commonly used in bioinformatics to predict the functional effect of a missense mutation. The authors also developed a publicly available Web-tool (Variant Effect Prediction) to estimate a consensus score taking into account the results from four different tools (SIFT, PolyPhen2, SNPs&GO and Mutation Assessor).
Since the automated prediction of functional impact is part of most SNV prioritization pipelines, this paper could certainly be useful to develop a robust NGS secondary analysis.

Genomics. 2013 Jul 3. pii: S0888-7543(13)00126-2
Predicting the functional consequences of non-synonymous DNA sequence variants - evaluation of bioinformatics tools and development of a consensus strategy.

Frousios K, Iliopoulos CS, Schlitt T, Simpson MA. 

Abstract

The study of DNA sequence variation has been transformed by recent advances in DNA sequencing technologies. Determination of the functional consequences of sequence variant alleles offers potential insight as to how genotype may influence phenotype. Even within protein coding regions of the genome, establishing the consequences of variation on gene and protein function is challenging and often requires substantial laboratory investigation. However, a series of bioinformatics tools have been developed to predict whether non-synonymous variants are neutral or disease-causing. In this study we evaluate the performance of nine such methods (SIFT, PolyPhen2, SNPs&GO, PhD-SNP, PANTHER, Mutation Assessor, MutPred, Condel and CAROL) and developed CoVEC (Consensus Variant Effect Classification), a tool that integrates the prediction results from four of these methods. We demonstrate that the CoVEC approach outperforms most individual methods and highlights the benefit of combining results from multiple tools.