Genome-wide delineation of natural variation for pod shatter resistance in Brassica napus

Abstract

Resistance to pod shattering (shatter resistance) is a target trait for global rapeseed (canola, Brassica napus L.), improvement programs to minimise grain loss in the mature standing crop, and during windrowing and mechanical harvest. We describe the genetic basis of natural variation for shatter resistance in B. napus and show that several quantitative trait loci (QTL) control this trait. To identify loci underlying shatter resistance, we used a novel genotyping-by-sequencing approach DArT-Seq. QTL analysis detected a total of 12 significant QTL on chromosomes A03, A07, A09, C03, C04, C06, and C08; which jointly account for approximately 57% of the genotypic variation in shatter resistance. Through Genome-Wide Association Studies, we show that a large number of loci, including those that are involved in shattering in Arabidopsis, account for variation in shatter resistance in diverse B. napus germplasm. Our results indicate that genetic diversity for shatter resistance genes in B. napus is limited; many of the genes that might control this trait were not included during the natural creation of this species, or were not retained during the domestication and selection process. We speculate that valuable diversity for this trait was lost during the natural creation of B. napus. To improve shatter resistance, breeders will need to target the introduction of useful alleles especially from genotypes of other related species of Brassica, such as those that we have identified.

Figure 1. Distribution of shatter resistance, as measured with the pendulum test, among DH lines from the BLN2762/Surpass 400 population grown under three environments: experiment 1 (2011, screenhouse, SHT11); experiment 2 (2012, screenhouse, SHT12); and experiment 3 (2012, screenhouse, SHTWW12).

Pair-plots of EBLUPS from DH lines and parental lines showing correlations are presented. Rupture energy (RE) was measured in mJ. Data from experiments SHT11, SHT12 and SHTWW12 were RE, whereas for experiments RELSQ11, RELSQ12 and RELSQWW12 the data were RE (adjusted for pod length) from the same lines.

Figure 2. Box-plot showing variation for resistance to pod shatter in population of DHs from BLN2762/Surpass400, grown under three different environments: Experiment 1, 2010, Birdcage; Experiment 2, 2011, Birdcage; and Experiment 3, 2011, Field conditions).

Figure 3. QTL detected with linkage (whole average genome interval mapping – WGAIM [DH-QTL], statistical machine learning-SML [DH-QTL], and genome-wide association analysis [GWAS] in Brassica napus germplasm.

Marker sequences were aligned with the sequenced genomes of B. rapa and B. oleracea and their physical positions are shown with dotted lines (Tables S1& S4). Putative candidate genes (marked with red lines) that were localised within the physical map intervals are listed. Only QTL consistent across environments are shown (Table 1): a = Qrps.wwai-A03a and Qrps.wwai-A03b; b = Qrps.wwai-C03; c = Qrps.wwai-A09a, and Qrps.wwai-A09; b and d = Qrps.wwai-C08a, Qrps.wwai-C08b and Qrps.wwai-C08c.

Figure 4. Anatomical differences among four exemplar haplotypes from the DH population derived from BLN2762/Surpass400.

a = DH line 6668 is a haplotype with Shp(B) and Ind(S) alleles, b = DH line 6823 is a haplotype with Shp(B) and Ind(B) alleles, c = DH line 7128 has Shp(S) and Ind(B) alleles, and d = DH line 7124 has Shp(S) and Ind(S) alleles. Alleles B and S given in parentheses represent the parental donor lines of the DH population. Arrows show marked structural differences among haplotypes.