Homeologous Exchange and Segregation Generate Genetic Diversity in Resynthesized Brassica napus for Ten Generations Following Allopolyploid Formation
Allopolyploids are typically produced by hybridization between species followed by whole genome duplication (WGD). Many plant species including a number of important crops are allopolyploids or are diploids with evidence of allopolyploidy in their evolution [1-7]. Researchers use resynthesized allopolyploids to study the genetic changes that accompany hybridization and WGD. The Brassica species represent a well-established example of allopolyploidy  and resynthesized Brassica allopolyploids have been useful tools for studying the genetic effects of allopolyploid formation . Hybridization of B. rapa (n=10, designated the “A” genome) and B. oleracea (n=9, the “C” genome) followed by WGD produces the allopolyploid B. napus (n=19, designated “CCAA”). As observed with other allopolyploids, B. napus displays heterosis (i.e. hybrid vigor) relative to the diploid progenitors. Resynthesized B. napus show evidence of chromosomal rearrangements and changes in DNA methylation beginning at the first meiosis following WGD (i.e. in the S0 generation)[12,13]. Not surprisingly, these genetic changes can be accompanied by phenotypic changes; lineages of resynthesized B. napus have been identified with changes in the S5 generation relative to the S0 in flowering time, stature, pollen viability and seed set. Genetic changes experienced by allopolyploids impact reproductive success and are, therefore, important from the perspective of plant evolution.
Previous studies of resynthesized B. napus have analyzed the genetic variation of plants from distinct lineages at the samegeneration (i.e. 50 distinct S0 or S5 plants are compared). These studies have effectively sampled the types of genetic changes experienced by resynthesized B. napus plants. The driver for most genetic changes is the high degree of synteny between the A and C genomes[15,16]. Homeologous pairing and recombination during meiosis in B. napus can shuffle the genome producing chromosomes bearing both A and C loci[13,17,18]. This homeologous exchange (HE) results in partial or complete aneuploidy in gametes resulting in progeny that vary from the expected AACC dosage at some loci (e.g. AAAC and AAAA dosages)[12,14,19]. The application of fluorescence in-situ hybridization(FISH) to B. napus has confirmed that HEs are frequent in resynthesized B. napus and revealed larger changes involving loss of an entire A or C chromosome. Chromosome loss can be unbalanced (AAC) or balanced (AAAC) depending on whether or not the loss of a chromosome is accompanied by the gain of a homeolog. Interestingly, while chromosome loss during meiosis appears to be frequent, levels of aneuploidy among resynthesized B. napus are fairly low with most plants having 36-42 chromosomes. This observation suggests that balanced dosage changes tend to predominate. HEs and chromosome dosage change are not unique to B. napus and areobserved in other allopolyploid species[22,23].
Shuffling of homeologous chromosomes is thought to contribute to the stabilization of the allopolyploid genome in a processknown as “diploidization” . As part of this process homeologous pairing becomes suppressed and the allopolyploidgenome begins to behave genetically as a diploid. It is thought that loss of some homeologous loci could drive thisprocess[19,23]. Here we analyze several lineages of resynthesized B. napus and determine that novel genotypes appearthroughout the first ten generations following hybridization and WGD. In these lineages 10 generations is not sufficient fordiploidization to have occurred suggesting that diploidization is either a more gradual process or that diploidization early inan allopolyploid lineage is a rare event unlikely to be observed in a small sampling of lineages.
Materials and Methods
Plant Breeding and Growth
Hybridization of Brassica oleracea (TO1000, egg donor; C-genome) and Brassica rapa (IMB218; pollen donor; A-genome)produced 50 resynthesized B. napus allopolyploid plants (CCAA) as described previously by Lukens et al.  and Gaeta et al. . Brassica rapa and B. oleracea are doubled haploids, and thus are expected to be homozygous at every locus. The CA hybrids produced in the original crosses were treated with colchicine to induce genome doubling resulting in the first allopolyploid generation (S0).
For three lineages (EL5, EL37, and EL78) five plants from each generation S1-S11 were grown in a growth chamber with a 16h:8h day:night cycle for phenotype analysis. For each line and each generation, leaves from five individuals were pooled for DNA extraction (described below). Two lineages (EL5 and EL78) that retained high fertility through the S10 generation were chosen for further study. Single S1, S6, and S11 plants from lineage EL5 and lineage EL78 were self-pollinated to generate populations of S2, S7 and S12 individuals (referred to as “sibling populations”). Three individuals in the EL5 S6 generation were self-pollinated to generate populations of 35 siblings each. These were named TW5-6-1, TW5-6-2, and TW5-6-3.
Seeds of the sibling populations were germinated in germination mix with 1.5 g Osmocote Plant Food fertilizer (The ScottsCompany, Marysville, CA, USA) at California Polytechnic State University, San Luis Obispo, CA, USA. Two weeks after germination, plants were transplanted to 2-gallon pots containing commercial planting mix and organized in a completely random design on an open roof of a building from June 2009 to September 2009. Plants were top-watered using a drip irrigation system with 1L/48 hrs. Plants were top-watered with 0.5 mL/L DynaGro™ fertilizer (Dyna-Gro, Richmond, CA, USA), applied to the surface of the soil every seven days once flowering began. All plants were sprayed with 0.005% Green Light Malathion Insect Spray (Green Light Company, Longview, TX, USA) every 30 days. Pollination bags were used to encourage self-pollination and prevent cross-pollination between individuals.
Young leaves were removed for DNA extraction at the time of flowering and frozen in liquid nitrogen then stored at -20°C.
DNA extraction and amplification
Genomic DNA was extracted from leaf tissue of plants using DNAeasy Plant Maxi Kit (Qiagen; Valencia, CA, USA). A collection of 23 polymorphic markers was used to assay sequence loss among the sibling populations. Markers used were BRMS006, BRMS008, BRMS016, BRMS040. Fito018, Fito036, Fito043, Fito068, Fito083, Fito095, Fito131, Fito139, Fito151, Fito476 . SN2834, SN11670, SR6688, SR7178, SR9555, SR12387, SS2066 (primer sequences for markers beginning SN, SR, SS downloaded from http://www.Brassica. info/resource/markers/ssr-exchange.php). In addition, polymorphic markers were generated for the FLC3 and FLC5 loci.FLC3 primers (FLC3f, 5’GGCTCATGCGCATTTAACGT 3’, FLC3r5’TAGGTGGACCAACAATGCAA 3’) amplify an HpaI restriction site present in B. rapa and altered by a single nucleotide in B.oleracea. FLC5 primers (FLC5f 5’TGTTTTCGTGTGCCAATGTT 3’, FLC5r 5’ GGTTTTAAACTGTTGTGCTCAT 3’) amplify an indel of approximately 20 bp. Fragments were amplified using the following conditions: 1.0 U of GoTaq Flexi DNA Polymerase(Promega Corporation; Madison, WI, USA), 2.25 mM MgCl2, 5X Green GoTaq Flexi Buffer, 2.5 μM of dNTPs, 10 μM each of forward and reverse primers, 10 ng of plant DNA, and dH2O to final volume of 20 μl. PCR conditions were as follows: 95°C for 30 sec; eight cycles of 94°C for 30 sec, 50°C for 30 sec, 72°C for 50 sec; 22 cycles of 89°C for 30 sec, 50°C for 30 sec, 72°C for 50 sec; 72°C for 30 sec. Each amplification included a negative control with 2 μl dH2O replacing plant DNA, and a positive control with a synthetic hybrid consisting of a 1:1 mixture of B. rapa and B. oleracea DNA. Fragments were separated using a 6.0% polyacrylamide gel for 50 min at 300V. Fragments were visualized using a UV transilluminator and analyzed using Quantity One 4.6.3 Gel Doc EQ (BioRad, Hercules, CA, USA). All B. napus samples that produced a fragment pattern different than the synthetic hybrid were replicated.
PCR and sequencing primers for pyrosequencing (Supplemental Table 1) were designed using PyroMark Assay Design Software 2.0 (Qiagen, USA). Pyrosequencing assay was developed by the PyroMark Q24 Analysis Software (Qiagen, USA) and performed on the PyroMark Q24 (Qiagen, USA).
For the FLC3 locus PCR mixture final concentrations used were: 0.06 U of GoTaq Flexi DNA Polymerase (Promega Corporation; Madison, WI, USA), 2.2 mM MgCl2, 1X Green GoTaq Flexi Buffer, 1.3 μM dNTPs, 0.2 μM each of forward and reverse primers, 10 ng of plant DNA, and dH2O to final volume of 25 μl. For the Bra012552 locus PCR mixture final concentrations used were: 0.06 U of GoTaq Flexi DNA Polymerase (Promega Corporation; Madison, WI, USA), 2.2 mM MgCl2, 1X Green GoTaq Flexi Buffer, 0.8 μM dNTPs, 0.2 μM each of forward and reverse primers, 10 ng of plant DNA, and dH2O to final volume of 25 μl. For the Bra017743 locus PCR mixture final concentrations used were: 0.06 U of GoTaq Flexi DNA Polymerase (Promega Corporation; Madison, WI, USA), 2.2 mM MgCl2, 1X Green GoTaq Flexi Buffer, 0.8 μM dNTPs, 1X of Qiagen custom oligos, 10 ng of plant DNA, and dH2O to final volume of 25μl.
Pyrosequencing Binding Mix for the FLC3, Bra012552, and Bra017743 loci was 40 μl Binding Buffer, 18 μl Nanopure water, 2 μl streptavidin beads. The assay buffer mix for FLC3 and Bra012552: 25 μl Annealing Buffer and 0.75 μl 10μM Sequencing Primer. The assay Buffer Mix for the Bra017743 SNP: 22.5 μl Annealing Buffer and 2.5 μl 10x Sequencing Primer.
For analysis of DNA fragment loss, the null hypothesis of equal proportion of genetic change was tested between generations within a lineage with analysis of variance (ANOVA) and between lineages with 2-sample t-tests. Chi-Square goodnessof- fit test was used to compare observed segregation patterns with Mendelian ratios.
Lineages of Brassica napus undergo phenotypic and genotypic changes over ten generations
Previously, 50 independent lineages of B. napus were generated using well-characterized diploid parents B. oleracea (TO1000; C subgenome) and B. rapa (IMB218; A subgenome) (figure 1a) [12,14]. Many of these lineages have been propagated to the 11th generation (S11) by single-seed descent. For three of these lineages (EL5, EL37, EL78), five plants were grown for each generation (S1-S11) and phenotypes were compared. One lineage assumed a dwarf phenotype at the S6 generation (Figure 1b,c). samples for each generation were analyzed with polymorphic markers capable of distinguishing the A and C subgenome. Among the three lineages, loss of either an A or C subgenome locus was detected at the S1, S3, S4, S5, S6 and S11 generation suggesting that genetic changes can occur throughout the early generations of a resynthesized lineage (Figure 1d).
Phenotypic variation and sequence loss observed among progeny of Brassica napus individuals
For two lineages (EL5 and EL78) individuals at the S1, S6, and S11 generations were self-pollinated to produce pools of 35S2, S7 and S12 progeny. Many novel phenotypes were observed within sibling populations (Figure 2). Plants with altered stature and flower color were self-pollinated and their progeny analyzed. The phenotypes were shown to be heritable in thenext generation confirming that these phenotypes reflect genetic changes (data not shown).
DNA was extracted from the parent and from approximately 35 individual progeny from each generation. Parents andprogeny were screened with 24 polymorphic microsatellite markers to identify individuals with loss of either C or A subgenome sequences at those loci. In every sibling population, individuals were identified that lacked sequences present in the parent. At the S2 generation the B. rapa and B. oleracea subgenomes lost marker sequences at the same frequency (2-sample t-test;α=00.5, P >0.05) while in the S7 and S12 generations marker sequences were more frequently lost from the B. oleracea subgenome (2-sample t-test;α=00.5, P=0.000 for S7 and S12 in both lineages).
Loss of a locus or loci that had been detected in the parent was observed in all six sibling populations. Between 3 to 18 individuals in each sibling population lacked at least one A or C subgenome locus that had been detected in the parent. Several distinct patterns of sequence loss were observed in the progeny populations. Five cases were observed in which the same marker signal was lost from ≥5 individuals in a population. In these 5 cases there was no significant difference between the observed segregation and the 3:1 ratio expected from Mendelian segregation (γ2 <2.70 for all). In two populations homelogous loci were lost from different siblings (i.e. one sibling lost the A genome locus while a different sibling lost the corresponding C genome locus). In one population, three individuals lost all loci analyzed on the A7 chromosome. In addition, six other individuals in that population lost one or more loci on the A7 chromosome.
Figure 1. Phenotype and genotype analysis of Brasscia napus lineages. A) Resynthesized B. napus (center) is the allopolyploid formed by hybridization of B. rapa (IMB218 (left)) and B. oleracea (TO1000 (right)). Bars represent 10 cm. B) Average height of the first flower for each generation S1-S11 for three lineages (n=5, error bars represent the standard deviation of the mean). c) Representative individuals from the EL37 lineage from generation S1 (left) through S11 (right). Bar represents 10 cm. D) Representative image of genotype analysis. Two lineages are analyzed with the same microsatellite marker (C, TO1000; SH, synthetic hybrid made by mixing equal concentrations of TO1000 and IMB218 genomic DNA; A, IMB218). Marker amplification from the S1-S10 generations for two lineages are shown. The lineage to the right shows loss of A subgenome signal at the S5 generation.
Figure 2. Novel phenotypes within a sibling population of B. napus. All plants pictured are from the same sibling population and are approximately the same age. A) Allopolyploid B. napus generated from TO1000 and IMB218 have white flowers. B) A yellow-flowered plant. C) A rosette-dwarf individual (bar represents 2 cm). D) Variation in flowering time. The plant on the left produced 5 leaves before flowering while the plant on the right produced 18 leaves before flowering.
Analysis of chromosomal changes with dosage-sensitive markers
A dosage-sensitive polymorphic marker was developed for the FLC3 locus and analyzed by pyrosequencing. Pyrosequencingallows identification of both balanced (CAAA and CCCA) and unbalanced (CAA and CCA) changes in dosage at this locus(Figure 3a).
Figure 3. Dosage-sensitive analysis of the FLC3 locus. A) Synthetic hybrids created by mixing DNA from the C and A genome parents analyzed by pyrosequencing. Error bars represent the standard deviation of the mean (n=4). B) Analysis of progeny of EL5 S2. C) Analysis of progeny of EL5 S6. In B and C, white circles represent the average values for controls. The size of the white circles is arbitrary and does not represent variation. Controls are C only (upper left), CCCA, CCA, CA, CAA, CAAA, A only (lower right). For progeny (black squares) points represent the mean of two replicates (see supplemental data for standard deviation of the mean). Arrows represent the FLC3 dosage detected for the parent.
Two sibling populations (EL5 at the S2 and S6 generations)were screened and in both cases, the parent was determined to have equal FLC3 dosage (CCAA). However, progeny of these two parents revealed distinct patterns of segregation and bothpopulations contained individuals with FLC3 dosage distinct from the parent. In one population (EL5 at S6) individuals wereidentified with equal dosage at the FLC3 locus (CCAA) and balanced changes toward either B. oleracea or B. rapa (CCCA orCAAA). Additionally, several individuals appear to have unbalanced dosage changes and three individuals were identified with loss of either C or A signal (figure 3c, Supplemental Figure1a). In the other population (EL5 at the S2) 32 of 35 individualsappear to have unchanged dosage (CCAA) and three individuals have balanced changes in dosage (CAAA or CCCA) (figure3b, Supplemental Figure 1b).
To further determine the effect of chromosome dosage on segregation three individuals from EL5 S6 progeny population(described above) were selected based on their A and C chromosome dosage at the FLC3 locus (see Materials and Methods for explanation of naming). These individuals had either CCAA (TW5-6-1), A only (TW5-6-2), or CAA (TW5-6-3) at the FLC3 locus. For these three plants their dosage was determined at two additional loci on the A3:C3 homeologs (Bra012552, Bra017743). The plants were self-pollinated to produce approximately 34 progeny. Dosage was determined at FLC3, Bra012552, and Bra017743. Among the 33 progeny of parent TW5-6-1, only 4 had novel genotypes at any of the loci tested (Figure 4a). Three of these four had a dosage change at the FLC3 locus while the other had a dosage change at locusBra012552. Parent TW5-6-2 also gave rise to progeny with few novel genotypes (Figure 4b). Only one of 33 progeny wereidentified with a novel dosage change (at the Bra017743 locus). Parent TW5-6-3 had an unbalanced dosage change at FLC3 (CAA) (Figure 4c) and gave rise to many progeny with novel dosages at the FLC3 and Bra012552 loci.
Figure 4. Dosage-Sensitive Analysis of Three Loci on the A3:C3 Homeologs. Grey bars represent the A3:C3 chromosome. Dark grey circles represent the A3:C3 centromere. Boxes show the approximate position of the FLC3, Bra021552, and Bra017743 loci. Color at each locus corresponds to the dosage detected. A) TW5-6-1 parent and 33 selfed progeny. B) TW5-6-2 parent and 33 selfed progeny. C) TW5-6-3 and 34 selfed progeny.
Figure 5. A Mendelian Model for Segregation of HEs. Homeologous chromosomes are distinguished by color (grey: B. oleracea (C subgenome), white: B. rapa (A subgenome)). The parent (Sn generation) has an non-reciprocal HE at the top of the homeologous chromosomes shown. Reproduction is through self-pollination. The genotype shown for each Sn+1 genotype reflects the dosage in the HE region. 25% of progeny are expected to lack A subgenome loci within the HE region. This model assumes that, despite the HE, the homologs shown in white will pair during meiosis resulting in each gamete receiving either the chromosome bearing the HE or the unaffected chromosome.
Allopolyploidy confers both advantages and challenges to plants. Many allopolyploids display hybrid vigor (heterosis)relative to the diploid progenitors. However, allopolyploids also experience genomic disruption and loss of loci through pairing and exchange between homeologous chromosomes [6,7,10]. Homeologous recombination can disrupt the production of euploid gametes and limit reproductive success . Here, we show that 11 generations after hybridization and WGD, resynthesized B. napus continues to undergo homeologous recombination resulting in novel genotypes. Additionally, segregation of homeologous rearrangements is a source of genetic variation in resynthesized populations of B. napus plants. Both shifts in dosage from the expected CCAA genotype and complete loss of loci from one progenitor or the other are detected.
Figure 6. Possible segregation in an individual with a balanced dosage change. C subgenome shown in grey, A subgenome shown in white. The parent (Sn) carries a balanced dosage change (CCCA). The sole A chromosome could be lost during meiosis or undergo homeologous pairing with a C chromosome. Homeologous recombination and segregation is proposed to result in many possible Sn+1 genotypes some of which are shown above. Genotypes lacking some or all A subgenome loci on this homeolog pair are boxed in bold.
Homeologous pairing in meiosis is observed in resynthesized allopolyploid plants [13,28-30] but most studies examine thegenomes of these plants in the generations immediately after resynthesis. This study, which takes advantage of longer lineages of resynthesized B. napus, expands the understanding of genomic instability in recently-formed allopolyploids and supports the idea that diploidization and suppression of homeologous exchange may develop gradually . When S2, S7 andS12 plants are compared within a single lineage novel genetic changes (i.e. the loss of either a C genome or A genome locus)were observed in every generation tested. While this study only examines two lineages in this way, the results support thenotion that homeologous pairing still occurs and that lineages are continuing to diverge genetically 11 generations afterhybridization and WGD.
Populations of siblings were observed in which multiple individuals lost the same locus either from the B. oleracea subgenome or the B. rapa subgenome. It is unlikely that multiple individuals in a relatively small population (n=32) wouldexperience loss of the same loci independently; therefore, these changes are attributed to segregation of a homeologous rearrangement present in the parent plant. We suggest that segregation of HEs provide an explanation for some of the loci lost from multiple individuals within a population. In one possible model a parent with a non-reciprocal HE produces offspringdeficient for the exchanged region at a frequency of 25% (Figure5). Five cases were identified in which the same loci was lost from 5 or more siblings and in all of these cases the ratio of locus-present:locus-absent did not differ significantly from a 3:1 ratio. This model assumes that the HE region is not so large that homologous pairing is disrupted (i.e. the two A genomehomologs will still pair faithfully during meiosis despite the presence of the exchanged segment of C subgenome present on one of the A subgenome chromosomes). HEs involving larger segments could drive homeologous pairing possibly leading to an increase in the number of progeny lacking loci from one subgenome.
Gaeta and Pires  suggest a segregation pattern resulting from segregation of a reciprocal HE. This model suggests a1:4:6:4:1 ratio of genotypes among the progeny of a plant with a reciprocal HE (CCCC : CCCA : CCAA : CAAA : AAAA). Analysis of one population at the FLC3 locus with a dosage-sensitive marker revealed a segregation pattern very similar to this prediction. It is not possible to detect balanced dosage changes with the techniques used here, however, the segregation pattern strongly suggests the presence of a balanced HE around the FLC3 locus in the parent.
Three individuals from one sibling population apparently lacked the A7 chromosome (no signal was detected from markers distributed across the A7 chromosome). Six different individuals in that population lacked the A subgenome locus for one, two, or three of the A7 markers. We speculate that the parent of this population (in which all A7 markers were detected) had a balanced dosage change (CCCA, Figure 6) in which one A7 chromosome is replaced by a homeologous C subgenomechromosome (either C7 which is syntenous with the top of A7, or C6 which is syntenous with the bottom half of A7 .
Duing meiosis in this parent the single A7 chromosome would, in the absence of a homolog, be lost or form a homeologous pair with a C chromosome. Homeologous recombination between the single A7 chromosome and a C chromosome could result in a relatively high portion of gametes lacking some interval of A7. Observing loss of C subgenome loci from an A7 homeolog is unlikely in this small population because most gametes from the parent are expected to contain one or more copies of every C locus. Thus, some gametes of a plant with a balanced dosage change are expected to have to either total loss or partial loss of the chromosome present in reduced dosage in the parent.
The observations described above and in previous studies that characterize genetic changes in allopolyploids have focused on the deletion of loci from one progenitor or the other. These PCR-based approaches are dosage-insensitive and, therefore, do not distinguish between balanced and unbalanced changes in the allopolyploid genome. Analysis of allopolyploid siblings using a dosage-sensitive detection method (pyrosequencing) reveals both de novo dosage changes in individuals and intriguing patterns of segregation resulting from existing dosage changes. Application of a dosage-sensitive marker for the FLC3 locus provided a way to distinguish between segregation of HEs present in the parent and de novo rearrangements that occur during meiosis in the parent. One sibling population (Figure 3b) included three individuals that had a change in dosage at the FLC3 locus (two were CCCA and one was CAAA) with the rest retaining the same CCAA dosage as the parent. A possible explanation of this observation is that the parent did not contain a HE around the FLC3 locus and that the individuals with CCCA or CAAA genotypes are the result of novel HEs occurring during meiosis in the parent. If this is the case then the rate of homeologous pairing and recombination must be quite high (at least at this locus) as 3 out of 70 meiosis events have appeared to have involved HE at this locus. FLC3 is located at the distal end of chromosome 3 in B. rapa and B. oleracea. The A3/C3 chromosomes are acrocentric and it is known from Arabidopsis that rates of recombination are elevated as distance from the centromere increases . Therefore, the frequency of HE near the FLC3 locus may be higher than that of loci closer to centromeres.
When progeny from three of the plants described above are analyzed with additional dosage-sensitive markers on the A3/C3 homeologs the dosage of the parent strongly influenced the frequency of novel dosage changes in the next generation. Appearance of novel dosage changes was rare in the progeny of parents with no dosage change or a balanced dosage change. However, the progeny of TW5-6-3, which had an unbalanced dosage at the FLC3 locus produced many progeny with novel dosages at FLC3 and at the other two loci tested. Clearly, this investigation does not represent a ‘chromosome-wide’ analysis as only three markers are used. We can’t conclude definitely that the dosage imbalance at the distal end of the chromosome in TW5-6-3 is the driver for the dosage changes observed at other locations in the TW5-6-3 progeny. However, the observations suggest that dosage imbalance at one loci predispose the homeologs to undergo additional HE during meiosis.
This work was funded by the NSF Plant Genome Program Award # 0733857, The California Polytechnic State University Extramural Funding Initiative, and by the Cal Poly Student Faculty Interaction Program. We are grateful to Patrick Edger at University of Missouri, Columbia, for providing Brassica genome sequence; We thank Jen Vanderkelen, Joshua Dillard for pyrosequencing training and advice. Samantha Fong, Kyle Fujimoto, Jenny McCarthy, and Jessica Burke assisted with plant growth and marker testing. We are grateful to Rebecca Deorge, Denise Bradford, Soma Roy, and John Walker for assistance with statistical analysis.
1. Ozkan H, Levy AA, Feldman M. Allopolyploidy-induced rapid genome evolution in the wheat (Aegilops-Triticum) group. Plant Cell. 2001,13(8): 1735-1747.
3. Lim KY, Soltis DE, Soltis PS, Tate J, Matyasek R et al. Rapid Chromosome Evolution in Recently Formed Polyploids in Tragopogon (Asteraceae). PLoS ONE. 2008, 3(10): e3353.
4. Pontes O, Neves N, Silva M, Lewis MS, Madlung A, Comai A, Viegas W, Pikaard CS. 2004. Chromosomal locus rearrangements are a rapid response to formation of the allotetraploid Arabidopsis suecica genome. Proc Natl Acad Sci U S A. 2004, 101(52): 18240-18245.
6. Lashermes P, Hueber Y, Combes M-C, Severac D, Dereeper A, 2016. Inter-genomic DNA Exchanges and Homoeologous Gene Silencing Shaped the Nascent Allopolyploid Coffee Genome(Coffea arabica L.) G3 g3.116.030858. doi:10.1534/ g3.116.030858.
9. Himelblau E, Lauffer D, Teutonico R, Pires JC, and Osborn TC. Rapid-cycling Brassica in research and education. In: Pua, Douglas, eds. Biotechnology in Agriculture and Forestry, Brassica Biotechnology, Berlin: Springer-Verlag. 2014, 54: 13-28.
12. Lukens LN, Pires JC, Leon E, Vogelzang R, Oslach L et al. Patterns of sequence loss and cytosine methylation within a population of newly resynthesized Brassica napus. Plant Physiology. 2006, 140(1): 336-348.
16. Iniguez-Luy FL, Voort AV, Osborn TC. 2008. Development of a set of public SSR markers derived from genomic sequence of a rapid cycling Brassica oleracea L. genotype. Theoretical and Applied Genetics. 2008, 117(6):997-985.
17. Nicolas SD, Leflon M, Monod H, Eber F, Coriton O et al. Genetic Regulation of Meiotic Cross-Overs between Related Genomes in Brassica napus Haploids and Hybrids. The Plant Cell. 2009, 21(2): 373-385.
18. Ecke W, Kampouridis A, Ziese-Kubon K. Hirsch AC. Identification and genetic characterization by high-throughput SNP analysis of intervarietal substitution lines of rapeseed (Brassica napus L.) with enhanced embryogenic potential. Theor Appl Genet. 2015, 128(4): 587-603.
21. Xiong Z, Gaeta RT, Pires JC. Homoeologous shuffling and chromosome compensation maintain genome balance in resynthesized allopolyploid Brassica napus. Proc Natl Acad Sci U S A. 2011, 108(19): 7908-7913.
23. Tate JA, Joshi P, Soltis KA, Soltis PS, Soltis DE. On the road to diploidization? Homoeolog loss in independently formed populations of the allopolyploid Tragopogon miscellus (Asteraceae). BMC Plant Biology. 2009, 9:80.
25. Suwabe K, Tsukazaki H, Iketani H, Hatakeyama K, Kondo M et al. Simple Sequence Repeat-Based Comparative Genomics Between Brassica rapa and Arabidopsis thaliana: The Genetic Origin of Clubroot Resistance. Genetics. 2006, 173(1): 309-319.
26. Iniguez-Luy FL, Lukens L, Farnham MW, Amasino RM, Osborn TC. 2009. Development of public immortal mapping populations, molecular markers and linkage maps for rapid cycling Brassica rapa and B. oleracea. Theor Appl Genet. 2009, 120(1): 31-43.
31. Jan Drouaud, Christine Camilleri, Pierre-Yves Bourguignon, Aurelie Canaguier, Aurelie Bérard et al. Variation in crossing-over rates across chromosome 4 of Arabidopsis thaliana reveals the presence of meiotic recombination “hot spots”. Genome Research 2006, 16(1): 106-114.