SNP discovery in complex allotetraploid genomes (Gossypium spp., Malvaceae) using genotyping by sequencing

Abstract

Premise of the study: Single-nucleotide polymorphism (SNP) marker discovery in plants with complex allotetraploid genomes is often confounded by the presence of homeologous loci (along with paralogous and orthologous loci). Here we present a strategy to filter for SNPs representing orthologous loci.

Methods and results: Using Illumina next-generation sequencing, 54 million reads were collected from restriction enzyme-digested DNA libraries of a diversity of Gossypium taxa. Loci with one to three SNPs were discovered using the Stacks software package, yielding 25,529 new cotton SNP combinations, including those that are polymorphic at both interspecific and intraspecific levels. Frequencies of predicted dual-homozygous (aa/bb) marker polymorphisms ranged from 6.7-11.6% of total shared fragments in intraspecific comparisons and from 15.0-16.4% in interspecific comparisons.

Conclusions: This resource provides dual-homozygous (aa/bb) marker polymorphisms. Both in silico and experimental validation efforts demonstrated that these markers are enriched for single orthologous loci that are homozygous for alternative alleles.

Keywords: Gossypium; genotyping by sequencing; interspecific; intraspecific; next-generation sequencing; polyploid; single-nucleotide polymorphisms.

Fig. 1.

Predicted marker type categories from the sstacks algorithm for four common genetic scenarios (out of many) that give rise to apparent GBS polymorphisms between two allotetraploids. Red lines indicate sequences that can be clearly assigned to the A_T subgenome, and blue lines indicate those that can be assigned to the D_T subgenome. Gray lines indicate regions of high sequence similarity between homeologs or paralogs (e.g., no differences outside of the SNP of interest). Marker type predictions are based on the assumption that there is adequate sequence coverage to accurately score all alleles at all relevant loci.

Fig. 2.

BsrG1 and HinP1I GBS polymorphism in tetraploid Gossypium spp. The proportion of highly informative (aa/bb) markers relative to total shared loci (stacks) in intraspecific and interspecific pairwise comparisons is shown. 3-79 = G. barbadense cv. Pima 3-79; K-56 = G. barbadense accession K-56; TM-1 = G. hirsutum cv. TM-1; TX-231 = G. hirsutum accession TX-231.

Fig. 3.

Representative examples of the five categories of sequence alignments observed in TM-1 vs. Pima 3-79 polymorphic markers with aa/bb marker type assignment from Stacks. Nucleotides on a black background indicate the site of the key Pima 3-79 polymorphism relative to the TM-1 reference sequence. Nucleotides on a gray background indicate additional mismatches relative to the TM-1 reference sequence. The top two lines in each category indicate the TM-1 and Pima 3-79 fragment sequences, respectively. The prefix B indicates BsrG1 markers, and H indicates HinP1I. Additional lines in the alignment represent fragments from diploid genomes along with chromosomal assignments. BGI_A = Gossypium arboreum (Li et al., 2014); JGI_D = G. raimondii (Paterson et al., 2012); BGI_D = G. raimondii (Wang et al., 2012); scaf = scaffold.