A sequence-based genetic linkage map as a reference for Brassica rapa pseudochromosome assembly

The genus Brassica is one of the core genera within the tribe Brassicea. It comprises a large number of crops with a wide spectrum of morphological variation that can be cultivated under a variety of agro-climatic conditions. Brassica s provide vegetable oil, fresh and preserved vegetables, fodder and condiments, as well as being important sources of dietary fibre, vitamin C and nutritionally beneficial factors including anti-cancer compounds [1]. There are six representative species in the Brassica genus including three diploid species B. rapa (AA, 2n = 20), B. nigra (BB, 2n = 16) and B. oleracea (CC, 2n = 18), and three amphidiploids B. juncea (AABB, 2n = 36), B. napus (AACC, 2n = 38) and B. carinata (BBCC, 2n = 34). The genetic relationships among these Brassica species are well defined in U's triangle [2]. One of the diploid species, B. rapa, comprises a variety of morphologically diverse cultivated types including Chinese cabbage, misuna, aburana, flowering cabbage, turnip, turnip rape, yellow sarson, tatsoi and komatsuna, and these provide leaf heads, leaves, flowering stems, turnips and seeds, the productive organs for economical consumption [3, 4]. Furthermore, B. rapa is an excellent model for studying polyploidy genome evolution owing to its paleohexaploid ancestry and its close evolutionary relationships with Arabidopsis thaliana[5].

The multi-national B. rapa Genome Sequencing Project (BrGSP) was launched in 2003 owing to the economical and biological importance of B. rapa, and the A3 chromosome was sequenced using traditional Sanger technology [5]. In 2008, rapid next generation sequencing technology was employed for B. rapa genome sequencing and a high density genetic map based on sequence-tagged markers is necessary to anchor the assembled scaffolds to chromosomes. Several maps concerning B. rapa have been published to be used as reference genetic maps for pseudochromosome assembly (http://www.brassica-rapa.org) [6–8] despite their coverage only allowing in total 73.6% of the scaffolds being anchored on to chromosomes [The Brassica rapa Genome Sequencing Project Consortium: The genome of the mesohexaploid crop species Brassica rapa, submitted].

Genetic mapping is important for understanding the origin of and relationships among the genomes of Brassica species. Genetic linkage maps can also provide improved insight into genome organization and evolution through comparative mapping, and serve as the basis for genetic studies concerning various agronomic traits through the localization of major genes and quantitative trait loci (QTLs). Furthermore, they can aid breeding programs with the development of marker assisted selection (MAS) [9]. More than 20 genetic linkage maps have been constructed for B. rapa using a range of marker types including Restriction Fragment Length Polymorphisms (RFLPs), random amplified polymorphic DNA (RAPD), amplified fragment length polymorphisms (AFLPs), sequence-related amplified polymorphisms (SRAPs) and simple sequence repeats (SSRs) [6–8, 10–20]. However, there are limited published data concerning sequence-tagged PCR-markers, predominantly SSRs, mapped in B. rapa[6–8, 15, 19], particularly markers that could provide anchors for the B. rapa genome that are transferable to other mapping populations.

Recent developments in sequencing technology have simplified and accelerated the discovery of sequence variants, enabling development of sequence-based markers including single nucleotide polymorphisms (SNPs) and insertion/deletion polymorphisms (InDels) [21]. InDels and SNPs are the markers of choice for high-resolution genetic mapping and association studies owing to their abundance and distribution throughout the genome [22, 23]. For example, a study investigating genetic variation on human chromosome 22 suggested that InDels represent 18% of the polymorphisms on this chromosome [24]. Studies concerning genetic variation in A. thaliana have demonstrated that InDels represent 34% of all genetic polymorphisms [25]. Furthermore, InDels can contribute directly to a phenotype [26], or can associate with a phenotype as a result of linkage disequilibrium [27]. By re-sequencing 1,398 sequence-tagged sites (STSs) in eight B. rapa genotypes, Park et al. identified and characterized 6,753 InDels in the gene space of the B. rapa genome [28]. InDel polymorphisms are the second most frequent type of polymorphism in the genome, and can be genotyped using simple procedures including the analysis of size polymorphisms of polymerase chain reaction (PCR) products on agarose gels [29, 30]. Another advantage of InDel markers is the improbability of two InDel mutations being exactly the same length and at the same genomic position. Therefore, shared InDels represent identity-by-descent [31]. Very few InDel markers have been used to construct genetic linkage maps of B. rapa. With the recent completion of the sequencing of the B. rapa genome [The Brassica rapa Genome Sequencing Project Consortium: The genome of the mesohexaploid crop species Brassica rapa, submitted], the development of whole genome-wide InDel markers based on re-sequencing has become feasible, and this will be a useful resource for the international research community.

In this study, InDel and SSR markers, both of which are sequence-tagged PCR markers, were used to construct a high resolution genetic map of B. rapa. The map was used as a reference linkage map to anchor and orient sequence scaffolds for B. rapa genome assembly.

This study concerns the comprehensive linkage analysis of B. rapa. The map spans 1234.2 cM and is divided into ten linkage groups corresponding to the number of B. rapa chromosomes, with an average distance of 2.43 cM between adjacent markers. A conspicuous characteristic of the present map is that 415 markers, accounting for 81.7% of the total mapped markers, are novel InDel markers, which increases the number of sequence-based markers for B. rapa.

One of the main purposes of this linkage map is to support the B. rapa genome sequencing project to anchor and orient scaffolds onto the chromosomes. Using publically available genetic linkage maps for B. rapa including VCS_DH (354 markers), JWF3P (498 markers) (http://www.brassica-rapa.org) [6] and CK_DH (313 markers and 719 markers) [7, 8], 84 (171 Mb), 97 (185 Mb) and 91 (175 Mb) scaffolds were anchored to chromosomes, respectively. However, selecting markers that were evenly distributed along the genome and complementary to the previously reported maps, using the RCZ16_DH map, 152 (235 Mb) scaffolds were anchored on to the chromosomes. The RCZ16_DH map alone covers more than 82.9% of the B. rapa genome, indicating the potential of this sequence-based linkage map for linkage and QTL analysis.

In B. rapa, polymorphic DNA loci are relatively frequent. Park et al. demonstrated that the frequency of SNPs and InDels were 15.3 SNPs/kb and 4.83 InDels/kb by resequencing 1,398 STSs based on the 557 BAC sequences of B. rapa ssp. pekinensis cv Chiifu-401-42 [28]. Among the 28,222 sequence variants of B. rapa, approximately 24% were InDels. This high frequency of InDels has also been reported for other species including maize [39, 40], sugar beet [41], barley [42] and Arabidopsis [25]. In the present study, PCR-based InDel markers were developed using 80 to 200 bp PCR products, and the insertion/deletion size varied between 4 to 10 bp. The InDel polymorphisms were genotyped using a simple procedure that analyzed size polymorphisms of PCR products.

The diploid Brassica genomes contain large replicated blocks of collinear segments within and between linkage groups. These are thought to have derived from a polyploid ancestor, although the exact mechanism by which this occurred is debatable [43]. In the linkage map generated herein, only three SSR markers detected multiple loci and no evidence of conserved blocks of synteny can be deduced from this. This low level of detection of the replication within the genome could be due to the marker types used to construct the map. The InDel markers were developed directly from scaffold sequences by selecting unique InDels, and ambiguous markers were excluded from the marker data set. Furthermore, SSR markers are usually located in non-coding sequences, which are less well conserved between replicated blocks than coding regions.

A total of 92 SSR markers are present in the map developed in this study, and 66 of these could be aligned to the scaffold sequences using stringent criterion of 100% match of the primer sequences. Among the 26 unmapped markers, 20 were designed on the basis of EST sequences of B. oleracea (BoE set and Ol set). For these markers, mismatches could exist within the primer sequences preventing them from being mapped to the scaffold sequence of B. rapa. Five of the six unmapped SSRs derived from B. rapa sequences were mapped to scaffolds with only one side primer, and this could be due to the gaps in the sequence assembly.

Alignment of the RCZ16_DH map to pseudochromosomes indicated high collinearity between the genetic and physical distance (Figure 2). However, the alignment result demonstrated that there were outliers distributed on the linkage groups A02, A04, A06, A08 and A10. This could be due to scaffolds being too small or the genomic region having too little recombination to allow precise placements or orientations, as a relatively small mapping population was used in the present study. The small population size could also lead to the relatively high statistical errors when there were missing marker data for some DH lines. The other possible reason for the inconsistency of genetic and physical distances could be assembling errors as the InDels were designed on the basis of the assembled scaffold sequences.

This study describes the use of sequence-based and highly polymorphic InDel markers to construct a highly resolute reference genetic map of B. rapa. The result is an improved resource for fine mapping of quantitative trait loci, identifying candidate genes and map-based gene isolation.