For rare or de novo CNVs, one might expect a potentially stronger effect on phenotype—but in a smaller number of samples, as is the case with the 16p11.2 deletion recently discovered (using Birdseye) that seems to explain 1% of idiopathic autism21, or one of numerous deletions associated with schizophrenia22. In such a scenario, aggregation of events in cases at a particular locus, coupled with a lack of events in controls, can indicate that the region affects the assayed phenotype. The new release of PLINK includes a set of tools for manipulating, summarizing and analyzing rare and de novo CNVs output from Birdseye (see URLs section in Methods).
