The CNV events identified by Birdseye, together with the Canary genotypes for common CNPs, yield an assessment of copy number for each sample across its genome; the SNP genotypes from Birdseed describe sequence variation for samples at SNPs with the expected two copies of each locus. The fourth component of Birdsuite, Fawkes (‘fast analysis with kopy-number et SNPs’), merges these results to yield an integrated picture of the genetic variation in each sample. Fawkes utilizes the imputed locations (in A/B intensity space) of copy-variable clusters to assign an allele-specific copy number genotype (such as AAB, ABBB, A or B) at each SNP (Fig. 1 and Supplementary Methods). Notably, the genotype assignment for a sample is constrained to the set of clusters corresponding to its integer copy number as determined by Canary and Birdseye; allelic copy number is thus informed by measurements not only from the SNP probe, but also from nearby SNP and copy number probes. This approach differs fundamentally from earlier attempts to estimate allele-specific copy number from intensity data at individual SNPs13,17.
