Previous approaches to copy number analysis7-10 involve searching a single individual’s genome for regions in which evidence of copy number deviation exceeds a genome-wide significance threshold—an approach that does not make use of prior knowledge. Yet the variation at more than 90% of the loci at which any two individuals differ in copy number across a region >10 kb in size seems due to a limited universe of common CNPs6. At such loci, a copy number variant unambiguously exists and segregates at an appreciable frequency, and the problem can be redefined not as a problem of ab initio discovery, but rather of accurate measurement (genotyping) of each individual’s integer copy number level5.
