From: Extending long-range phasing and haplotype library imputation algorithms to large and heterogeneous datasets
