Fig. 6

Iterative mapping of long-reads: Haplotype-specific references for haplotypes a and b are created iteratively. While some ambiguities remain in the long-read pileup after an initial round of mapping (mapIter1), alignment accuracy is significantly improved by remapping the reads sets against consensus sequences derived from the first mapping (mapIter2). The majority of HPs in the mapping are resolved at the end of the iterative mapping (lower panel). Some positions still harbour deletions, denoted by the purple bars, which need to be resolved by short-reads in the final mapping step