paperKB
coga / coga-kb
Help
Sign in

Chunk #50 — Online Methods — Step 1: Direct IBD-based phasing using long IBD

Source
Fast and accurate long-range phasing in a UK Biobank cohort.
Embedded
yes

Text

For each proband in turn, Eagle scans all other (diploid) individuals for long genomic segments (>4cM) in which one (haploid) chromosome is likely to be shared IBD with the proband. Eagle then analyzes these probable IBD matches for consistency, identifies a consistent subset, and uses this subset to make phase calls. In our N≈150,000 analyses, this step required ≈10% of the total computation time (Supplementary Table 2) and achieved near-perfect phasing within long swaths of genome covering most of each sample (corresponding to regions with IBD to several relatives) (Fig. 1a). In more detail, our algorithm applies the following four procedures to each proband in turn.