In this algorithm, each study individual gains phasing information from all other individuals in the dataset, which can lead to very accurate haplotype estimates at typed SNPs when the total sample size is large. Once a study individual has sampled a new pair of haplotypes, the imputation step is broken into two parts: SNPs in U2 are imputed using information from both the haploid and diploid reference panels (Step 4), and SNPs in U1 are imputed using only the haploid reference panel (Step 5). This modeling choice highlights a core principle of our inference framework: we allow the method to naturally adapt to the amount of information in the data by conditioning only on observed genotypes, not imputed ones, at each step.
