VerySNP: VCF feature based SVM to reduce false positive rate in SNP predictor output

Several open-source tools have been recently developed to identify Single Nucleotide Polymorphisms (SNPs) in whole-genome data, the most popular being SAMtools and GATK. Commonly, SNP predictors provide a VCF file as output, which contains a list of candidate SNPs and additional information such as SNP call quality and depth of coverage. Still, the SNP list presents an unsatisfactory accuracy due to high false positive polymorphism prediction. VCF parameters have been used to train a Support Vector Machine (SVM) that classifies the VCF SNP list in true and false positive SNPs, cleaning the SNP predictor output from the most likely false positive results. We implemented the SVM approach in a new software, called VerySNP, and applied it to model and non-model organisms proving, in both cases, this machine learning method efficiently recognizes true positive from false positive SNPs.