Calling Rare Variants from Genotype Data

Abstract

DNA microarrays have been successfully used for human genetics research to assay single nucleotide polymorphisms (SNPs) throughout the genome. DNA microarrays work by measuring the relative amount of binding of input DNA to a set of complementary oligonucleotide probes for each allele using a photometric assay. Once the raw data are collected, they need to be converted into a genotype call automatically and with high accuracy. Over the past decade, many groups have published calling algorithms that are able to achieve greater than 99.5 % accuracy. However, these algorithms work best for common SNPs and are not as accurate for low-frequency and rare variants (minor allele frequency <5 %). With the widespread usage of microarrays targeting rare variants such as Exome Chip and MetaboChip, new calling algorithms that accurately call rare variants have been published over the last year. In this chapter, we will describe how DNA microarrays work (see section “Microarray Technology”), give a brief overview of genotype calling algorithms (see section “Genotype Calling Algorithms”), and summarize the different algorithms designed for rare variants and how well they perform (see section “Application to Rare Variants”).