Mapping-by-sequencing in complex polyploid genomes using genic sequence capture: a case study to map yellow rust resistance in hexaploid wheat

Abstract

Previously we extended the utility of mapping-by-sequencing by combining it with sequence capture and mapping sequence data to pseudo-chromosomes that were organized using wheat-Brachypodium synteny. This, with a bespoke haplotyping algorithm, enabled us to map the flowering time locus in the diploid wheat Triticum monococcum L. identifying a set of deleted genes (Gardiner et al., 2014). Here, we develop this combination of gene enrichment and sliding window mapping-by-synteny analysis to map the Yr6 locus for yellow stripe rust resistance in hexaploid wheat. A 110 MB NimbleGen capture probe set was used to enrich and sequence a doubled haploid mapping population of hexaploid wheat derived from an Avalon and Cadenza cross. The Yr6 locus was identified by mapping to the POPSEQ chromosomal pseudomolecules using a bespoke pipeline and algorithm (Chapman et al., 2015). Furthermore the same locus was identified using newly developed pseudo-chromosome sequences as a mapping reference that are based on the genic sequence used for sequence enrichment. The pseudo-chromosomes allow us to demonstrate the application of mapping-by-sequencing to even poorly defined polyploidy genomes where chromosomes are incomplete and sub-genome assemblies are collapsed. This analysis uniquely enabled us to: compare wheat genome annotations; identify the Yr6 locus - defining a smaller genic region than was previously possible; associate the interval with one wheat sub-genome and increase the density of SNP markers associated. Finally, we built the pipeline in iPlant, making it a user-friendly community resource for phenotype mapping.

Keywords: genomics; mapping-by-sequencing; next generation; target enrichment; wheat.

Figure 1

Demonstrating the reorganization of capture probe design contigs between two pseudo‐genome assemblies. Highlighting how contig positions in our previous pseudo‐genome assembly that was generated using 807 Brachypodium markers (Gardiner et al., 2014), are translated, using connecting ribbons, to their respective positions in the pseudo‐genome that was generated here using 11 016 wheat markers from the Genome Zipper (Nussbaumer et al., 2013; Spannagl et al., 2013). Chromosomes are labeled according to the markers that were used to generate them; in the previous pseudo‐genome they are labeled Brachy 1–7 and in the newer pseudo‐genome they are labeled IWGSC 1–7. Chromosomes are colour coded and connecting ribbons show the colour of the chromosome of origin of the contig in the previous pseudo‐genome (Brachy 1–7).

Figure 2

Processing three sets of enriched sequencing data to identify a mapping interval containing the gene that is inducing the phenotype of interest. (a) Standard mapping and SNP calling pipeline to construct ‘reference genomes’. (b) Pipeline implementing an algorithm to score regions of interest by prioritizing long homozygous parental haplotypes for the bulk segregant sample to identify the interval of interest. Default figures shown here of 50 000‐bp windows at 1000‐bp intervals and standard homozygote/heterozygote/borderline SNP definitions shown‐values are adjusted throughout the analysis as necessary.

Figure 3

Homozygosity scores calculated for the bulk segregant dataset along each pseudo‐chromosome. Magenta: Scores plotted for ‘Cadenza specific homoeologous homozygote SNP alleles’ found in the bulk segregant dataset. Blue: Scores plotted for ‘Avalon‐specific homoeologous homozygote SNP alleles’ found in the bulk‐segregated dataset. Scores calculated per 500 000‐bp window along each chromosome at 10 000‐bp intervals.

Figure 4

Analysis of the peak intervals on pseudo‐chromosome 7. (a, c) Magenta line: Scores plotted for ‘Cadenza specific homoeologous homozygote SNP alleles’ found in the bulk segregant dataset. Scores calculated per 500 000‐bp window along each chromosome at 10 000‐bp intervals and displayed only for (a) the interval 13 650 001–14 150 001 bp on the MIPS‐derived pseudo‐chromosome 7 and (c) the interval 7 650 001–8 150 001 bp on the POPSEQ‐derived pseudo‐chromosome 7. (b, d) Uses the same x‐axis as plots (a, c) respectively to depict the same interval: symbols (|) highlight the positions of candidate SNPs that are highly homozygous in the bulk segregant dataset and conserved with the Cadenza parent, those coloured magenta show SNPs that overlap disease‐resistance‐related genes.

Figure 5

POPSEQ versus Genome Zipper genetic markers. Detailing 6043 IWGSC CSS contigs that have been assigned a genetic position (cM) by the Genome Zipper and also in the POPSEQ analysis. Contigs included only if both the Genome Zipper and POPSEQ analyses assign them to the same sub‐genome and chromosome.