Fig. 3

Genetic mapping, haplotype-specific assembly and validation. a Top: Genetic map with a total genetic length of 622.0 cM (see the “Online methods” section). Middle: up to 2-Gb reads were assigned to one of the two haplotypes of each linkage group. Bottom: a combination of haplotype-A/B linkage groups led to two assemblies with 214.6 and 215.3 Mb. b Contig size distributions before (ctg-A, ctg-B) and after scaffolding (scaf-CU for the assembly with sequence from “Currot”; and scaf-OR for the assembly with sequence from “Orange Red”). After scaffolding, eight chromosome-scale pseudo-molecules were obtained for each haplotype as labeled by “Chrs”. c Haplotype validation for the two assemblies of each linkage group (LG-1-8) using parent-specific k-mers (of “Orange Red” and “Currot”). With each linkage group, the two assemblies could be clearly identified as either “Currot” haplotype or “Orange Red” haplotype using parental k-mers. After combining the “Currot”-related assemblies and “Orange Red”-related assemblies to genome-level, k-mer comparison revealed a haplotype accuracy of 99.1%. d Using the “Currot” haplotype as representative and comparing it to the assembly of the double haploid Prunus ssp. reference genome (Prunus persica, and other closely related species; Additional file 1: Fig. S8) revealed high levels of synteny and thus implies high accuracy of the genetic map and chromosome-level scaffolding. e Hi-C contact map based on the assemblies of the two haplotypes of chromosome 1 (Currot (CU) and Orange Red (OR)) at a resolution of 300 kb. The contact signal showed a high contiguity within the haplotypes (main diagonal line) and confirmed two large inversions (v₁₁ and v₁₂) which we observed in the assembled sequence of the two haplotypes. (See Additional file 1: Figs. S9–15 for chromosomes 2–8)