The phasing and imputation calculations are driven by the hidden Markov model (HMM) of Li and Stephens (2003). This model can be used to update an individual's haplotypes by constructing them as “imperfect mosaic” copies of a set of template haplotypes. In the simplest case, the templates would include all reference+study haplotypes (minus the pair being updated) in Step 1 and all reference haplotypes in Step 2.
