Polyploidy or whole genome duplication (WGD) has occurred multiple times throughout the evolutionary history of plants. It has long been recognized as a major force in angiosperm evolution, plant speciation and diversification [1–3]. Polyploidization is both an ancient and an ongoing evolutionary process [4, 5] and has played a role in the adaptation of a wide range of crops to different environments by generating phenotypic variation. Polyploids are generally divided into two categories; autopolyploids from duplication of the same genome; and allopolyploids from hybridization of two diverged genomes with subsequent genome duplication. These distinctions are less clear in paleopolyploids. Soybean  and maize  are considered to be paleopolyploids having been formed between 10 – 15 Mya. Both show evidence of diploidization, an ongoing process by which a newly formed polyploid becomes stabilized, involving the loss of duplicated genes, thereby returning the genome to a diploid-like form . Both potato and alfalfa are derived through autopolyploidy, while wheat, oat, cotton, coffee and oilseed rape have allopolyploidy in their evolutionary history. Trapogoyon is a young allopolyploid species that has formed multiple times over the last 80 years  and so offers the opportunity to study a natural allopolyploid which is sympatric with its parental species . The success of newly formed angiosperm polyploids is partly attributable to their highly plastic genome structure. Recent studies have documented rapid and dynamic changes in genomic structure and gene expression in plant polyploids. Much of the functional plasticity in polyploids is correlated with gene expression changes at transcriptional and post-transcriptional levels. Such gene expression changes are controlled largely by epigenetic mechanisms [1, 2, 11].
The Brassica species include an important group of vegetable and oil crops and their genomes have complex evolutionary histories. A major focus for research has been Brassica napus (oilseed rape). This is an allopolyploid species formed by the hybridization of progenitor species Brassica rapa (which contributed the A genome) and Brassica oleracea (which contributed the C genome). The Brassica species in general, and B. napus in particular, provide an excellent system in which to study the impacts of polyploidy and the processes by which genomes subsequently stabilize. B. rapa and B. oleracea are closely related, having diverged around 3.5 Mya . The B. napus types cultivated as crops arose from natural polyploid formation, probably during human cultivation, i.e. less than 10,000 years ago. Genetic mapping studies confirmed that the progenitor A and C genomes are essentially intact in natural lines of B. napus and have not been substantially rearranged . It is also possible to make newly constructed (“resynthesised”) polyploids in the laboratory by crossing B. rapa and B. oleracea accessions and doubling chromosomes (typically by chemical treatment). Song et al. used resynthesised polyploids to study genome evolution in the early generations after polyploidization and demonstrated that polyploid species can generate extensive genetic diversity in a short period of time. Pires et al. were interested in the ability of polyploids to possess novel traits that are not present in their diploid progenitors which has allowed polyploids to successfully enter new ecological niches. Focussing on flowering time they showed evidence of chromosomal rearrangements and changes in gene expression, which partially explained the phenotypic variation in B. napus. The mechanisms for chromosome stability and diploidization in polyploids remain largely unknown but a study of 50 resynthesised lines of B. napus showed that in the first generation (S 0 ) of resynthesised B. napus, genetic changes are rare but cytosine methylation changes are frequent, whereas in later generations (S 5 ) genetic changes are much more frequent, but the S 0 methylation remained fixed . The genetic changes observed in resynthesised B. napus are not random and there is evidence that many are the consequence of homoeologous recombination . Recent cytological investigations including a S 10:11 generation showed that changes in copy number of individual chromosomes increased with successive generations; they showed gross chromosomal rearrangements and that dosage balance mechanisms enforced chromosome number stability . There is much interest on how these genetic and epigenetic changes contribute to changes in gene expression. Transcriptional changes are likely to be a critical component of polyploid evolution as they can contribute directly to novel phenotypes. Most studies have compared gene expression in resynthesised polyploid lines to expression in their parents to provide evidence of additive or non-additive gene expression. According to the "additivity hypothesis”, newly-synthesized allopolyploids are supposed to display mid-parental expression patterns. Many exceptions are found in resynthesised allopolyploids e.g. Arabidopsis, Senecio, Brassica, Triticum, and Gossypium, suggesting that the differential regulation of gene expression is a common feature of plant allopolyploids. Although the phenomenon of non-additive expression in inter-specific hybrids and allopolyploids is now well described, the underlying mechanisms are still poorly understood. Recent studies have used statistical methods to predict the contribution of each parent to gene expression in the polyploid using genome-wide microarrays that are not able to distinguish between expression of homoeologous pairs [17, 22]. The “additivity” hypothesis was confirmed using comparative proteomics on newly resynthesised B. napus. Identification using mass spectrometry and functional categorisation of the differentially regulated proteins did not show that any functional category, metabolic pathway or subcellular localization was over- or under represented within non-additive polypeptides . Comparing transcript levels in resynthesised B. napus to protein levels showed that differential protein regulation is not explained by transcriptional changes . This is a complex process so another approach has been to measure transcript levels of homoeologous pairs of genes, but not transcriptome-wide. For example, Dong et al. showed a complex pattern of differential expression in response to abiotic stress in both natural and resynthesised allopolyploid Gossypium hirsutum using SSCP-cDNA gels to distinguish homoeologous pairs of 60 genes. Also, Chaudhary et al. used a mass-spectrometry-based SNP detection technique to measure allele- and homoeologue-specific contributions to the transcriptome of diploid and allopolyploid cotton and showed that 40% of homoeologues were transcriptionally biased in at least one stage of cotton development. Development of a method to measure genome-wide differential expression of homoeologous pairs using transcriptome sequencing in both synthetic and natural polyploids would contribute to our understanding of this complex process.
Next generation sequencing technologies (NGS) have opened exciting opportunities to study genomes and transcriptomes of plant species with and without sequenced genomes. Many crop genome projects are ongoing, including oilseed rape, bread wheat and banana, but many of these polyploid plants have complex genome structures meaning that producing a draft sequence is challenging [28, 29]. Meanwhile plant transcriptomics using NGS can yield much information on crops , including gene discovery, transcript quantification, post-transcriptional regulation and linking genotypes to phenotypes . mRNA-Seq is a recently developed approach to transcriptome profiling that uses deep-sequencing technologies . Previous experiments, using mRNA-Seq for SNP detection in B. napus[33, 34], have proved Illumina sequencing to be an efficient method. mRNA-Seq can also be used as a method to estimate transcript abundance. The first step is to map the reads to the genome or transcriptome, and then the number of reads aligning to a specific region of the reference sequence is counted and subjected to relevant normalisation procedures . It is anticipated that mRNA-Seq will revolutionize the manner in which eukaryotic transcriptomes are analysed  as sequencing-based approaches have clear advantages over hybridization-based approaches for quantifying the transcriptome. A range of studies comparing microarray and mRNA-Seq have consistently shown that sequencing has higher sensitivity and dynamic range [36, 37], although reproducibility has been shown to depend on the type of sample studied [36, 38, 39] and a recent study has shown that technical variability is too high to be ignored .
We previously developed a set of 94,558 Brassica unigenes, by assembly of all available Brassica ESTs, and used this set for the design of a Brassica microarray . We used this for the analysis of gene expression in resynthesised B. napus lines (“B. napus 1” and “B. napus 2”). These resynthesised B. napus lines shared the same parental combinations of B. rapa (R-o-18) and B. oleracea (A12DH), but from reciprocal crosses. The Brassica unigenes were assembled using parameters that enabled the separate assembly of transcripts of paralogous genes within each diploid Brassica genome (as these differ by ~15% at the nucleotide level) but the co-assembly of transcripts of homoeologous genes (which differ by only ~3%). The surface-bound oligonucleotide probes of the microarrays were designed to regions that differed most between unigenes (typically 3’ untranslated regions), so discriminate well between unigenes representing paralogous genes, but they have no capability to discriminate between homoeologous genes.
In the present study, we report the development of methodology for using mRNA-Seq to quantify transcript abundance in polyploids, with estimation of the relative contributions of homoeologous genes. As a proof of concept, we analysed mRNA from reserved aliquots of the same ground leaf samples taken from the resynthesised B. napus plants used in our 2009 microarray-based study. We show, inter alia, that mRNA-Seq can be used successfully for both qualitative and quantitative analyses of gene expression and for the apportioning of transcript abundance to A and C genomes in both resynthesised and natural B. napus.