Fig. 4

Comparison of Hi-C based phasing and scaffolding with gamete binning. a Hi-C contact map based on all 16 haplotype assemblies generated with gamete binning (Currot (CU) and Orange Red (OR)). b Hi-C contact map based on all 16 haplotype assemblies generated with Hi-C data (Currot (CU) and Orange Red (OR)). Note the contact signal along the main diagonal was much weaker as compared to the signal based on the gamete binning assembly, and virtually no contact signals between two different haplotypes could be identified. c Haplotype validation of each linkage group (LG-1-8) of the Hi-C assembly using parent-specific k-mers. Almost all haplotype-specific assemblies included k-mers specific to both of the parental alleles indicating severe errors in the phasing. d Alignment of the Hi-C based assembly to genetic map derived assembly (i.e., gamete binning derived assembly) revealed mis-joining or splitting of linkage groups within the Hi-C assembly. For example, CUR1G was split into Hi-C:19 g1, 19 g4; on the other hand, alignments of 19 g2 to CUR2G, CUR5G, and CUR7G revealed mis-joins of sequences from independent linkage groups