De novo SNP discovery and genetic linkage mapping in poplar using restriction site associated DNA and whole-genome sequencing technologies

Abstract

BackgroundRestriction site associated DNA sequencing (RAD-seq), a next-generation sequencing technology, has greatly facilitated genetic linkage mapping studies in outbred species. RAD-seq is capable of discovering thousands of genetic markers for linkage mapping across many individuals, and can be applied in species with or without a reference genome. Although several analytical tools are available for RAD-seq data, alternative strategies are necessary for improving the marker quality and hence the genetic mapping accuracy.ResultsWe demonstrate a strategy for constructing dense genetic linkage maps in hybrid forest trees by combining RAD-seq and whole-genome sequencing technologies. We performed RAD-seq of 150 progeny and whole-genome sequencing of the two parents in an F1 hybrid population of Populus deltoides × P. simonii. Two rough references were assembled from the whole-genome sequencing reads of the two parents separately. Based on the parental reference sequences, 3442 high-quality single nucleotide polymorphisms (SNPs) were identified that segregate in the ratio of 1:1. The maternal linkage map of P. deltoides was constructed with 2012 SNPs, containing 19 linkage groups and spanning 4067.16 cM of the genome with an average distance of 2.04 cM between adjacent markers, while the male map of P. simonii consisted of 1430 SNPs and the same number of linkage groups with a total length of 4356.04 cM and an average interval distance of 3.09 cM. Collinearity between the parental linkage maps and the reference genome of P. trichocarpa was also investigated. Compared with the result on the basis of the existing reference genome, our strategy identified more high-quality SNPs and generated parental linkage groups that nicely match the karyotype of Populus.ConclusionsThe strategy of simultaneously using RAD and whole-genome sequencing technologies can be applied to constructing high-density genetic maps in forest trees regardless of whether a reference genome exists. The two parental linkage maps constructed here provide more accurate genetic resources for unraveling quantitative trait loci and accelerating molecular breeding programs, as well as for comparative genomics in Populus.Electronic supplementary materialThe online version of this article (doi:10.1186/s12864-016-3003-9) contains supplementary material, which is available to authorized users.

Construction of High-quality Linkage Maps in Populus with RADseq Data and their Application in QTL Mapping and Genome Assembly

“…trichocarpa [7]. In contrast, Mousavi et al [30] proposed a strategy for the de novo generation of SNPs with RADseq data across 150 progeny in an F 1 hybrid population of P. deltoides P. simoniii, by incorporating whole-genome sequencing data from the two parents. This strategy successfully identified 3,442 high-quality SNPs and produced the two parental linkage groups under a wide range of LOD (logarithm of odds) thresholds, which perfectly matched the karyotype of Populus.…”

“…To perform genetic linkage mapping and QTL analysis in Populus, we established an F 1 hybrid population by crossing P. deltoides and P. simoniii over three years from 2009 to 2011. The female parent, P. deltoides, and the male parent, P. simoniii, exhibit substantial differences in their growth rates and stress responses, thus producing permanent material for genetic mapping [30]. In our previous studies [29][30][31], different strategies for performing genetic mapping with RADseq data from the F 1 population were investigated, ultimately leading to the development of the novel software gmRAD for the de novo extraction of high-quality SNP genotypes.…”

“…The female parent, P. deltoides, and the male parent, P. simoniii, exhibit substantial differences in their growth rates and stress responses, thus producing permanent material for genetic mapping [30]. In our previous studies [29][30][31], different strategies for performing genetic mapping with RADseq data from the F 1 population were investigated, ultimately leading to the development of the novel software gmRAD for the de novo extraction of high-quality SNP genotypes. In the present study, we incorporated more individuals from the F 1 hybrid population to construct two parental high-quality genetic maps, and gmRAD was utilized for the generation of high-quality SNP genotypes solely from RADseq data.…”

Accounting for Errors in Low Coverage High-Throughput Sequencing Data when Constructing Genetic Maps using Biparental Outcrossed Populations

“…Typically, genotyping errors resulting from low sequencing coverage are removed via filtering, such as setting genotypes with an associated read depth below some threshold value to missing (Gardner et al . 2014; Mousavi et al . 2016) or using genotype quality scores to discard uncertain genotype calls (Mousavi et al .…”

“…2014; Mousavi et al . 2016) or using genotype quality scores to discard uncertain genotype calls (Mousavi et al . 2016; Hyma et al .…”