Making the Call

MassGenomic's Dan Koboldt discusses his latest Bioinformatics paper. In it, he presents VarScan, which is a program he developed to call SNPs and indels from next-gen sequencing data. The two main challenges to making accurate calls are telling the difference between real variants and false positives in short reads and holding the volume of data. "VarScan detects sequence variants, combines them by position and type, and then computes the read counts, average base quality, and number of strands supporting each allele," Kolbodt says. He later adds that most SNPs called by VarScan were in dbSNP, supported by Illumina/Solexa calls, or both.