Re-annotation of the physical map of Glycine max for polyploid-like regions by BAC end sequence driven whole genome shotgun read assembly.

Many of the world's most important food crops have either polyploid genomes or homeologous regions derived from segmental shuffling following polyploid formation. The soybean (Glycine max) genome has been shown to be composed of approximately four thousand short interspersed homeologous regions with 1, 2 or 4 copies per haploid genome by RFLP analysis, microsatellite anchors to BACs and by contigs formed from BAC fingerprints. Despite these similar regions,, the genome has been sequenced by whole genome shotgun sequence (WGS). Here the aim was to use BAC end sequences (BES) derived from three minimum tile paths (MTP) to examine the extent and homogeneity of polyploid-like regions within contigs and the extent of correlation between the polyploid-like regions inferred from fingerprinting and the polyploid-like sequences inferred from WGS matches. Results show that when sequence divergence was 1-10%, the copy number of homeologous regions could be identified from sequence variation in WGS reads overlapping BES. Homeolog sequence variants (HSVs) were single nucleotide polymorphisms (SNPs; 89%) and single nucleotide indels (SNIs 10%). Larger indels were rare but present (1%). Simulations that had predicted fingerprints of homeologous regions could be separated when divergence exceeded 2% were shown to be false. We show that a 5-10% sequence divergence is necessary to separate homeologs by fingerprinting. BES compared to WGS traces showed polyploid-like regions with less than 1% sequence divergence exist at 2.3% of the locations assayed. The use of HSVs like SNPs and SNIs to characterize BACs wil improve contig building methods. The implications for bioinformatic and functional annotation of polyploid and paleopolyploid genomes show that a combined approach of BAC fingerprint based physical maps, WGS sequence and HSV-based partitioning of BAC clones from homeologous regions to separate contigs will allow reliable de-convolution and positioning of sequence scaffolds (see BES_scaffolds section of SoyGD). This approach will assist genome annotation for paleopolyploid and true polyploid genomes such as soybean and many important cereal and fruit crops.

Saini N ，Shultz J ，Lightfoot DA 《BMC GENOMICS》

Gene duplication and paleopolyploidy in soybean and the implications for whole genome sequencing.

Soybean, Glycine max (L.) Merr., is a well documented paleopolyploid. What remains relatively under characterized is the level of sequence identity in retained homeologous regions of the genome. Recently, the Department of Energy Joint Genome Institute and United States Department of Agriculture jointly announced the sequencing of the soybean genome. One of the initial concerns is to what extent sequence identity in homeologous regions would have on whole genome shotgun sequence assembly. Seventeen BACs representing approximately 2.03 Mb were sequenced as representative potential homeologous regions from the soybean genome. Genetic mapping of each BAC shows that 11 of the 20 chromosomes are represented. Sequence comparisons between homeologous BACs shows that the soybean genome is a mosaic of retained paleopolyploid regions. Some regions appear to be highly conserved while other regions have diverged significantly. Large-scale "batch" reassembly of all 17 BACs combined showed that even the most homeologous BACs with upwards of 95% sequence identity resolve into their respective homeologous sequences. Potential assembly errors were generated by tandemly duplicated pentatricopeptide repeat containing genes and long simple sequence repeats. Analysis of a whole-genome shotgun assembly of 80,000 randomly chosen JGI-DOE sequence traces reveals some new soybean-specific repeat sequences. This analysis investigated both the structure of the paleopolyploid soybean genome and the potential effects retained homeology will have on assembling the whole genome shotgun sequence. Based upon these results, homeologous regions similar to those characterized here will not cause major assembly issues.

Schlueter JA ，Lin JY ，Schlueter SD ，Vasylenko-Sanders IF ，Deshpande S ，Yi J ，O'Bleness M ，Roe BA ，Nelson RT ，Scheffler BE ，Jackson SA ，Shoemaker RC ... -  《-》

The Soybean Genome Database (SoyGD): a browser for display of duplicated, polyploid, regions and sequence tagged sites on the integrated physical and genetic maps of Glycine max.

Genomes that have been highly conserved following increases in ploidy (by duplication or hybridization) like Glycine max (soybean) present challenges during genome analysis. At http://soybeangenome.siu.edu the Soybean Genome Database (SoyGD) genome browser has, since 2002, integrated and served the publicly available soybean physical map, bacterial artificial chromosome (BAC) fingerprint database and genetic map associated genomic data. The browser shows both build 3 and build 4 contiguous sets of clones (contigs) of the soybean physical map. Build 4 consisted of 2854 contigs that encompassed 1.05 Gb and 404 high-quality DNA markers that anchored 742 contigs. Many DNA markers anchored sets of 2-8 different contigs. Each contig in the set represented a homologous region of related sequences. GBrowse was adapted to show sets of homologous contigs at all potential anchor points, spread laterally and prevented from overlapping. About 8064 minimum tiling path (MTP2) clones provided 13,473 BAC end sequences (BES) to decorate the physical map. Analyses of BES placed 2111 gene models, 40 marker anchors and 1053 new microsatellite markers on the map. Estimated sequence tag probes from 201 low-copy gene families located 613 paralogs. The genome browser portal showed each data type as a separate track. Tetraploid, octoploid, diploid and homologous regions are shown clearly in relation to an integrated genetic and physical map.

Shultz JL ，Kurunam D ，Shopinski K ，Iqbal MJ ，Kazi S ，Zobrist K ，Bashir R ，Yaegashi S ，Lavu N ，Afzal AJ ，Yesudas CR ，Kassem MA ，Wu C ，Zhang HB ，Town CD ，Meksem K ，Lightfoot DA ... -  《-》

Feasibility of physical map construction from fingerprinted bacterial artificial chromosome libraries of polyploid plant species.

The presence of closely related genomes in polyploid species makes the assembly of total genomic sequence from shotgun sequence reads produced by the current sequencing platforms exceedingly difficult, if not impossible. Genomes of polyploid species could be sequenced following the ordered-clone sequencing approach employing contigs of bacterial artificial chromosome (BAC) clones and BAC-based physical maps. Although BAC contigs can currently be constructed for virtually any diploid organism with the SNaPshot high-information-content-fingerprinting (HICF) technology, it is currently unknown if this is also true for polyploid species. It is possible that BAC clones from orthologous regions of homoeologous chromosomes would share numerous restriction fragments and be therefore included into common contigs. Because of this and other concerns, physical mapping utilizing the SNaPshot HICF of BAC libraries of polyploid species has not been pursued and the possibility of doing so has not been assessed. The sole exception has been in common wheat, an allohexaploid in which it is possible to construct single-chromosome or single-chromosome-arm BAC libraries from DNA of flow-sorted chromosomes and bypass the obstacles created by polyploidy. The potential of the SNaPshot HICF technology for physical mapping of polyploid plants utilizing global BAC libraries was evaluated by assembling contigs of fingerprinted clones in an in silico merged BAC library composed of single-chromosome libraries of two wheat homoeologous chromosome arms, 3AS and 3DS, and complete chromosome 3B. Because the chromosome arm origin of each clone was known, it was possible to estimate the fidelity of contig assembly. On average 97.78% or more clones, depending on the library, were from a single chromosome arm. A large portion of the remaining clones was shown to be library contamination from other chromosomes, a feature that is unavoidable during the construction of single-chromosome BAC libraries. The negligibly low level of incorporation of clones from homoeologous chromosome arms into a contig during contig assembly suggested that it is feasible to construct contigs and physical maps using global BAC libraries of wheat and almost certainly also of other plant polyploid species with genome sizes comparable to that of wheat. Because of the high purity of the resulting assembled contigs, they can be directly used for genome sequencing. It is currently unknown but possible that equally good BAC contigs can be also constructed for polyploid species containing smaller, more gene-rich genomes.

Luo MC ，Ma Y ，You FM ，Anderson OD ，Kopecký D ，Simková H ，Safár J ，Dolezel J ，Gill B ，McGuire PE ，Dvorak J ... -  《-》

The development of BAC-end sequence-based microsatellite markers and placement in the physical and genetic maps of soybean.

The composite map of soybean shared among Soybase, LIS and SoyGD (March 2006) contained 3,073 DNA markers in the "Locus" class. Among the markers were 1,019 class I microsatellite markers with 2-3 bp simple sequence repeats (SSRs) of >10 iterations (BARC-SSR markers). However, there were few class II SSRs (2-5 bp repeats with <10 iterations; mostly SIUC-Satt markers). The aims here were to increase the number of classes I and II SSR markers and to integrate bacterial artificial chromosome (BAC) clones onto the soybean physical map using the markers. Used was 10 Mb of BAC-end sequence (BES) derived from 13,473 reads from 7,050 clones constituting minimum tile path 2 of the soybean physical map ( http://www.soybeangenome.siu.edu ; SoyGD). Identified were 1,053 1-6 bp motif, repeat sequences, 333 from class I (>10 repeats) and 720 from class II (<10 repeats). Potential markers were shown on the MTP_SSR track at Gbrowse. Primers were designed as 20-24 bp oligomers that had Tm of 55 +/- 1 C that would generate 100-500 bp amplicons. About 853 useful primer pairs were established. Motifs were not randomly distributed with biases toward AT rich motifs. Strong biases against the GC motif and all tetra-nucleotide repeats were found. The markers discovered were useful. Among the first 135 targeted for use in genetic map improvement about 60% of class II markers and 75% of class I markers were polymorphic among on the parents of four recombinant inbred line (RIL) populations. Many of the BES-based SSRs were located on the soybean genetic map in regions with few BARC-SSR markers. Therefore, BES-based SSRs represent useful tools for genetic map development in soybean. New members of a consortium to map the markers in additional populations are invited.

Shultz JL ，Kazi S ，Bashir R ，Afzal JA ，Lightfoot DA ... -  《-》