Fig. 1

Workflow: Top panel: a The short-reads are mapped to the generic reference and a consensus is called to serve as a sample-specific reference. Short-reads (b) and long-reads (c) are mapped to the sample-specific reference. Left panel: Heterozygous positions (HPs) are defined in the short-reads (d) and used to infer the genotype at these positions in the long-reads (e). “True” HPs are distinguished from artefacts by linkage analysis (f) and only true HPs are clustered into haplotypes (g). Right panel: Haplotype-specific long-reads are used to iteratively define draft haplotype consensus sequences (h) which are polished by the short-reads to call final haplotype consensus sequences (i)