![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
|
Published online before print
June 7, 2006, 10.1101/gr.4708406 Genome Res. 16:934-946, 2006 ©2006 by Cold Spring Harbor Laboratory Press; ISSN 1088-9051/06 $5.00
Letter Following tetraploidy in an Arabidopsis ancestor, genes were removed preferentially from one homeolog leaving clusters enriched in dose-sensitive genes1 College of Natural Resources, University of California–Berkeley, Berkeley, California 94720, USA; 2 Department of Environmental Science, Policy & Management, University of California–Berkeley, Berkeley, California 94720, USA; 3 Department of Plant & Microbial Biology, University of California–Berkeley, Berkeley, California 94720, USA
Approximately 90% of Arabidopsis’ unique gene content is found in syntenic blocks that were formed during the most recent whole-genome duplication. Within these blocks, 28.6% of the genes have a retained pair; the remaining genes have been lost from one of the homeologs. We create a minimized genome by condensing local duplications to one gene, removing transposons, and including only genes within blocks defined by retained pairs. We use a moving average of retained and non-retained genes to find clusters of retention and then identify the types of genes that appear in clusters at frequencies above expectations. Significant clusters of retention exist for almost all chromosomal segments. Detailed alignments show that, for 85% of the genome, one homeolog was preferentially (1.6x) targeted for fractionation. This homeolog fractionation bias suggests an epigenetic mechanism. We find that islands of retention contain "connected genes," those genes predicted—by the gene balance hypothesis—to be resistant to removal because the products they encode interact with other products in a dose-sensitive manner, creating a web of dependency. Gene families that are overrepresented in clusters include those encoding components of the proteasome/protein modification complexes, signal transduction machinery, ribosomes, and transcription factor complexes. Gene pair fractionation following polyploidy or segmental duplication leaves a genome enriched for "connected" genes. These clusters of duplicate genes may help explain the evolutionary origin of coregulated chromosomal regions and new functional modules.
There is strong evidence for a past of tetraploidy or near tetraploidy in eukaryotic genomes (Vision et al. 2000  ; McLysaght et al. 2002; Vandepoele et al. 2003), including yeast (Wolfe and Shields 1997) and all flowering plant genomes (Vision et al. 2000; Blanc et al. 2000, 2003; Bowers et al. 2003; Maere et al. 2005; Paterson et al. 2006). The general evolutionary process of gene duplication, occasional retention, and subsequent divergence to new function has much case support (Lewis 1951; Ohno 1970; Li 1997; Kellis et al. 2004), although the typical fate of any gene duplicate is loss (Haldane 1933; Lynch and Force 2000). Arabidopsis thaliana, a dicot flowering plant (125-Mb genome with  26,000 annotated genes) (The Arabidopsis Genome Initiative 2000), has retained duplicates in overlapping syntenous blocks that have been explained by multiple tetraploidy events: Bowers et al. (2003) suggested three events, the youngest denoted by  . Vision et al. (2000) suggested five waves of segmental duplications rather than tetraploidies. Using synonymous nucleotide substitution rate data, Maere et al. (2005) called the most recent event "3R," and dated it considerably more recently than did Bowers and coworkers. The most recent event is likely to have been a whole-genome duplication and not segmental because the comparative gene tree approach (Chapman et al. 2004) used by Bowers and coworkers covers 80% of the Arabidopsis genome once only. Because we begin with the unique retained pairs list provided by Bowers et al. (2003), we retain their nomenclature in referring to the most recent tetraploidy event as .
Genes classified by different gene ontology (GO) categories are retained at different rates following tetraploidy. In yeast, Papp et al. (2003)
According to the gene balance hypothesis, a gene displays dosage effects increasingly as the subunit–subunit interactions of its product increase (protein quaternary structural complexity), or from interactions with products downstream in a regulatory cascade, particularly through the interaction of positive and negative regulatory effectors (Veitia 2002
Following a duplication event, fractionation back toward the nonduplicated state is known to occur (Lockton and Gaut 2005
Synteny Viewer Assembly Version 5 of the annotation data for all five chromosomes in Arabidopsis was extracted from TIGR (http://www.tigr.org/tdb/e2k1/ath1). These data were parsed into a MySQL database using custom Perl language scripts. Visual inspection of syntenic regions was accomplished using a software tool we developed called Synteny Viewer. A public version of this Viewer is available at http://synteny.cnr.berkeley.edu/AtCNS. The Synteny Viewer software is a series of databases and Perl scripts that produce and display dynamic images of syntenic pairs and their BLAST high scoring pairs (HSPs). The images are displayed via a Web browser, wherein each object on the image is a link that displays information about that object, such as genomic location, sequence, orientation, GenBank annotation, EST support, GO product designations, structural information, and so on. The syntenic chromosomal pairs are compared using bl2seq (Tatusova and Madden 1999 ), with parameters as defined by Inada et al. (2003) in a window that is usually ±10 kb around an -gene pair. The Viewer presents further alignment, secondary structure, and RNA expression tools to expedite analyses of DNA sequences. We used each of 3822 pairs provided by Bowers et al. (2003) to anchor the Viewer to syntenous regions and observed the patterns of retained sequences in the flanking chromosomal areas. High scoring pairs (HSPs) displayed with version 5 TIGR model annotation were initially observed in a 20-kb window (40 kb if homeologs were oriented +/–) and expanded to 100 kb or until synteny was visualized. Figure 1 is a typical screenshot of a cluster of retained genes anchored by -pair AON075: AT2G17640–AT4G35640. Using this tool, it became visually apparent that retained genes were clustered.
Refining gene pairs Bowers et al. (2003) published a list of gene pairs retained from the most recent tetraploidy event. This work used the version 3 GenBank annotation. The gene list is particularly useful because the authors’ investigation used a comparative phylogenetic gene tree approach to discriminate three stages of tetraploidy (we are using pairs generated by the most recent event only) and because they organized their list into 26 syntenous blocks and several smaller regions. Subsequent to the publication of their gene list, version 5 gene annotations replaced version 3. We manually proofed at least 20 kb of sequence surrounding each of the 3822 Bowers gene pairs using our Viewer. Bowers and coworkers called some, but not all, possible gene pairs within clusters of locally duplicate genes. For the purposes of this analysis, we used our Viewer to manually condense duplications, invalidating all but one among local duplicates. We considered a cluster of duplicate genes to be broken if three or more genes intervened or if the duplicate genes aligned poorly (<50% coverage of exons using bl2seq with a mismatch penalty of –1) and were also more identical to an additional -gene. If no such "better alignment" was available, a sequence domain match was enough to condense a local duplicate into a single duplicate gene space—a situation that was very rare. We invalidated very few genes as being unalignable. Additionally, we found some new gene pairs in which both members of the pair were annotated but missed on the Bowers list, in which only one gene had annotation, or in which both genes lacked any annotation; we added these new pairs as "Our Additional" (designated OA -pairs). In all, we condensed the original 3822 pairs to 3178 (Supplemental material 1). Of the 129 OA gene pairs, about half were actually on the Bower’s list but off by a few genes, or did not translate perfectly from version 3 to version 5. In 33 cases involving annotation error, a gene is used for defining two pairs rather than one (Supplemental material 1, Columns G and I).
Condensing the genome
Preparing an
The 22,209 genes that remain are referred to as the minimized genome, which is one measure of a total genome that includes all of the paired regions from the most recent tetraploidy event. These regions comprise
Bowers et al. (2003)
Clustering statistics Test for randomness (global) The purpose of this test is to determine if the pattern of gene retention within the minimized genome is random. We begin with the null hypothesis that the pattern of retention is random (even though Viewer observations indicated otherwise). Each of the syntenic blocks was represented as a string of 1s and 0s, 1 for retained, 0 for non-retained. This binary sequence was then divided into non-overlapping bins of 10 genes. The mean value of each bin was calculated and stored. If the mean value in each bin is, on average, close to the mean rate of retention for the region, the sequence is random. However, if many bins have a high mean value and the others have a low mean value, this indicates clustering of retention, and the sequence is not random. The randomness of the sequence was tested for each block with the G-test (Sokal and Rohlf 1995 ) using the calculated mean values for all bins and the expected mean value, which is the retention rate for the entire -region being surveyed. The G-statistic is then compared to a  2 table to determine the probability that the sequence is random.
Test for clustering (local)
Aligning
Were it not for the 33 cases of double gene usage (Methods; Supplemental material 1, Column G), the two homeologs of a syntenic region would have identical numbers of retained genes (1s) but different numbers of non-retained genes (0s). Knowing this, we wrote a Perl script to align the 1s and distribute the 0s for each region. Because the number of non-retained genes on one
The alignment starts with the first retained gene on the a homeolog and the first retained gene on the b homeolog. For example, if between the first and second retained genes there are four non-retained genes on a (100001) and one non-retained on b (101) existed between the first and second retained genes, we introduce three gaps (___) on the b homeolog to maintain the alignment. Since it was not possible to know where the non-retained (singlet) genes were positioned relative to each homeolog, the 0s were lumped together at the left-most (lower-numbered) boundary of the gap. Therefore, the beginning of the alignment will be 100001 for the a homeolog, and 10___1 for the b homeolog. This alignment process continued to the end of the
We estimated the significance of fractionation bias by calculating the ratio of non-retained genes in the a and b homeologs using the side with the greater value as the numerator so that the ratio is always >1. We first reduced any run of non-retained genes longer than 20 down to 20. This was necessary to apply the same gaps rules to both homeologs; the original methods of Bowers et al. (2003)
Gene Ontology
Clustering: Genome-wide Figure 2 shows the results of running a moving average (80-gene window size) over the entire genome (Supplemental material 1, Column A), including local duplications and transposons, plotting average frequency of retained genes (Supplemental material 1, Column E) on the y-axis. Note that the centromeric regions (shaded) are either void or very low in retained genes.
Alignment diagrams for homeologs and fractionation bias Figure 3 shows the alignment diagrams for a representative three of the 26 larger and eight smaller (SO) -regions identified by Bowers and coworkers. Two of these example alignments illustrate a surprising result: One of the homeologs has lost significantly more genes than the other during the process of fractionating the tetraploid back toward the diploid. Figure 3 A14, -region 14, exemplifies a region where fractionation is significantly biased (Table 2A). Note that much of the white space, gaps inserted to maintain this alignment (Methods), is on one of the homeologs—a sign of biased fractionation. Fractionated genes (gray) are now singletons, and retained genes are the blue pairs. Detailed inspection of the alignment diagram of Figure 3 A14 found no indication that the preferred homeolog switched from one to the other chromosome, as one might expect if fractionation rates were set soon after tetraploidy, and then limited homeologous chromosomal recombination occurred. The red/green bar at the top of A14 shows the breakpoints for four small inversions. Following the procedure of Bowers et al. (2003), we inverted the red regions on the larger-numbered (b) homeolog in order to obtain the syntenous A14 block (Methods).
The two additional exemplary alignments in Figure 3, A23 and A13, should be read exactly as was A14. A23 is a typical -region showing fractionation bias. A13 exemplifies the infrequent -region not displaying a significant fractionation bias. Supplemental material 2 shows alignment diagrams like these for all 26 larger and eight smaller (SO) -regions. In general, simply looking at an alignment finds regions where genes have been repositioned closer together, into clusters, on the overfractionated homeolog. Ovals in Figure 3 represent examples of clusters. The inversion breakpoints, where red touches green, are not associated with fractionation bias. If there had been homeologous recombination after bias was set, then biased chromosomal stretches would get switched around. The thin crossing-over line in Figure 3 A13 illustrates such imaginary switchpoints, but we do not consider such switchpoints in our analyses (Discussion).
Significance and extent of fractionation bias
Table 1 gives total gene counts for each Table 2B estimates fractionation bias using gap size (measured in genes) as the unit of fractionation. Gaps devoid of genes (represented by white space in our diagrams) are necessary to permit homeolog alignment. Gaps of gene size 2–10 are all significantly different from 1:1 (Table 2B). Gaps of one gene are not significantly biased for unknown reasons, and gaps from 11 to 20 genes are sometimes biased significantly and sometimes not. We conclude that fractionation bias is a consequence of many smaller gap sizes, whereas the rare larger gaps do not influence fractionation bias significantly.
Local clustering
Genes found preferentially in clusters of retention Some gene ontology (GO) categories appear in clusters of retention more often than expected based on their overall frequency of retention. Each GO term appearing in our minimized genome is represented in Figure 5 as a point on a plot of number of genes versus number of genes in retained clusters; the lines indicate 95% confidence limits around the linear regression line. The 40 GO terms above the 95% confidence interval are those that appear in clusters more often than expected based on their appearance in -regions. A "y" is used in the last column of Table 3 to denote that the GO category was significantly overrepresented in clusters of retained genes.
There are 274 GO categories in Arabidopsis with at least 40 genes. We refer to the ratio of frequency of retention in clusters to the frequency of genome-wide retention as the representation in clusters. A high value of that ratio indicates that a particular GO category appears in retained clusters more often than expected based on its level of retention throughout the genome. Those ratios ranged from 0.15 to 0. There were 14 GO categories with a representation in clusters >0.08 (Table 3), and seven of these GO terms were significantly clustered (from Fig. 5, denoted "y" in Table 3, last column). The middle segment of Table 3 lists those 23 GO categories with representation in clusters between 0.06 and 0.08 that were judged to be significantly overpositioned in retained clusters. Of the 40 significant GO terms, 30 are in Table 3. (We removed seven general terms, two terms with near-identical gene content to a similar term, and terms with too few genes.) The last segment of Table 3 includes 10 GO categories with lower representation in clusters.
There are particular GO categories commonly found in over-retained clusters, and likewise, there are particular GO categories absent from clusters (Table 3). The well-populated categories "transcription factor activity" (and other transcription-related) and "kinase activity" (and other signal-transduction-related) appear in clusters more frequently than expected in the whole genome. The most extreme and also significant category of Table 3 is GO: "ubiquitin conjugating enzyme activity." Although this GO term only includes seven genes in retained clusters, there are six additional terms related to "ubiquitin" in the Top 14. The ubiquitin-proteasome pathway for specific protein degradation is known to have a particularly complicated subunit structure (Goldberg 2003
-pairs are clusteredWe have shown that those genes not lost following the most recent tetraploidy in Arabidopsis are frequently retained in clusters (Table 1; Figs. 3, 4; Supplemental material 2 and 3). The pattern of retention is nonrandom for 49 of 52 homeologs as a whole (Table 1), and moving averages identify significant local clusters on almost all homeologs (Table 1; Fig. 4; Supplemental material 3). Biased fractionation is one explanation for such clusters, as readily seen in our detailed alignment graphics (Fig. 3, where obvious clusters are identified; Supplemental material 2). One homeolog, probably representing one or the other of the original parents of the tetraploid, has experienced more (1.6x on average) gene loss than its partner. The data of Table 2A indicate that 21 of 26 larger -regions and six of eight smaller (SO) -regions are significantly biased. This gives an -genome (the 26 larger regions) bias coverage of 85%, which approximates an entire genome. This biased fractionation brings retained genes together into clusters on the overfractionated homeolog (see Fig. 3). If fractionation were not biased, then retained genes would still be brought closer together during fractionation, but not into clusters that did not already exist before fractionation. In summary, we show that retained genes are clustered, and have found a mechanism that naturally generates such clusters. Table 2B shows that smaller gaps of two to 10 genes account for the fractionation bias. These supporting data are important because they show that larger gaps, such as might result from large deletions of >11 genes on one homeolog only, are not the explanation for fractionation bias. Closer examination of the alignment diagrams for those few apparently "random" regions, like A13 of Figure 3 and A10 of Supplemental material 2, indicates that there are occasionally gene-orientation switchpoints that mask bias. Such switchpoints might be expected if differential homeolog mutability (bias) were established immediately after tetraploidy, and some homeologous recombination followed for a few generations. If we included the possibility of such recombination, near 100% coverage could be argued, but the validity of the argument would be difficult to test. Fractionation bias switchpoints in our alignment data weakly support the hypothesis that there was a transient period of homeologous recombination following tetraploidy. However, this period must have been short, if it existed at all, because homeologs are not significantly scrambled.
Categories of genes in over-retained clusters
GO categories with genes that are preferentially retained after the most recent tetraploidy in Arabidopsis have been identified by previous workers, as reviewed above. Interestingly, GO terms that appear in clusters (Table 3) include these same genes preferentially retained after the
Fractionation bias requires that homeologs be differentially marked for epigenetic inheritance prior to fractionation
Fractionation bias was unexpected. That this bias is so uniformly evident after all these years is remarkable. Given the observed fractionation bias, differential homeolog mutability does generally fit into a history of nonadditive phenomena observed in hybrid and polyploid organisms. Gene expression in hybrid plants is not the sum of the gene expression of the parents, there is methylation of the underexpressed homeolog, and underexpression may often be reversed with chemicals known to demethylate DNA (Heslop-Harrison 1990
Lippman and co-workers (2004)
The Arabidopsis tetraploidy is in a phylogenetic void of sequenced genomes. On the other hand, the sequenced Baker’s yeast (Saccharomyces cerevisiae) genome carries the syntenous gene pairs of an ancient tetraploidy for which there are multiple sequenced, diploid yeast out-groups, and has two post-tetraploid sister species, Candida glabrata and Saccharomyces castellii. Scannell and co-workers (2006)
Mechanisms of fractionation
The mechanisms of fractionation are not necessarily those that operate during duplicate gene divergence. There is adequate evidence from a variety of organisms that a pair of genes, once retained in some way, will diverge in function either by subfunctionalization or neofunctionalization (Gu et al. 2002
Subfunctionalization has provided a popular explanation for duplicate retention (Force et al. 1999
Mechanisms explaining clusters of retained genes
There are at least three explanations for duplicate retention that evoke mechanisms that are in place before or at the time of tetraploidy. Each of them involves selection for the status quo established before the new tetraploid existed by not fractionating certain loci. If the fractionation mechanism were long deletions, then those loci resistant to fractionation would also protect nearby genes. Candidates for such deletion-resistant loci are: (1) chromosome cis integrity regions; (2) heterotic allotetraploid homeologous gene pairs; and/or (3) duplicate genes that are particularly susceptible to a haploinsufficiency phenotype (predicted by the gene balance hypothesis). Matrix attachment regions (MARs), examples of the first explanation, have been reported as a large fraction of mammalian phylogenetic footprints that are not gene-associated (Glazko et al. 2003 For any of these mechanisms to explain clustering of over-retained genes using the concept of linkage disequilibrium, they must operate in an environment where long deletions happen, and we have not proved that long deletions happen during fractionation. Biased fractionation explains clustering without any mechanical assumptions. A deletion-resistant gene could potentially boost the retention rate of a neighboring gene that would otherwise be lost. The test of this hypothesis is confounded, however, largely because long deletions cannot be proven as a fractionation mechanism and by the paucity of genes in the most specific GO categories. In summary, the molecular mechanism of fractionation remains unknown.
Repeated tetraploidies and gene content evolution
There is a tendency in all studied eukaryotes for coregulated genes to be linked on the same chromosomal region (including in Arabidopsis) (Williams and Bowles 2004
We thank Eric Lyons, Lakshmi Rapaka, Damon Lisch, Keith Slotkin, Maggie Woodhouse, and George Theodoris for useful criticism. This work was funded by USA NSF DBI#0337083 and NSF DBI#0349737.
4 Corresponding author.
E-mail freeling{at}nature.berkeley.edu; fax (510) 642-4995. [Supplemental material is available online at www.genome.org.] Article published online before print. Article and publication date are at http://www.genome.org/cgi/doi/10.1101/gr.4708406
Adams K.L., Percifield R., Wendel J.F. 2004. Organ-specific silencing of duplicated genes in a newly synthesized cotton allotetraploid. Genetics 168: 2217–2226. . The Arabidopsis Genome Initiative. 2000. Analysis of the genome sequence of the flowering plant Arabidopsis thaliana.. Nature 408: 796–815.[CrossRef][Medline] Avramova Z., Tikhonov A., Chen M., Bennetzen J.L. 1998. Matrix attachment regions and structural collinearity in the genomes of two grass species. Nucleic Acids Res. 26: 761–767. Birchler J.A., Riddle N.C., Auger D.L., Veitia R.A. 2005. Dosage balance in gene regulation: Biological implications. Trends Genet. 21: 219–226.[CrossRef][Medline] Blanc G. and Wolfe K.H. 2004. Widespread paleopolyploidy in model plant species inferred from age distributions of duplicate genes. Plant Cell 16: 1667–1678. Blanc G., Barakat A., Guyot R., Cooke R., Delseny M. 2000. Extensive duplication and reshuffling in the Arabidopsisgenome. Plant Cell 12: 1093–1101. Blanc G., Hokamp K., Wolfe K.H. 2003. A recent polyploidy superimposed on older large-scale duplications in the Arabidopsis thaliana.. Genome Res. 13: 137–144. Bowers J.E., Chapman B.A., Rong J., Paterson A.H. 2003. Unravelling angiosperm genome evolution by phylogenetic analysis of chromosomal duplication events. Nature 422: 433–438.[CrossRef][Medline] Casneuf T., De Bodt S., Raes J., Maere S., de Van Peer Y. 2006. Nonrandom divergence of gene expression following gene and genome duplications in the flowering plant Arabidopsis thaliana.. Genome Biol. 7: R13.[CrossRef][Medline] Chapman B.A., Bowers J.E., Schultz S.R., Paterson A.H. 2004. A comparative phylogenetic approach for dating whole genome duplication events. Bioinformatics 20: 180–185. Chapman B.A., Bowers J.E., Feltus F.A., Paterson A.H. 2006. Buffering of crucial functions by paleologous duplicated genes may contribute cyclicality to angiosperm genome duplication. Proc. Natl. Acad. Sci. 103: 2730–2735. Conant G.C. and Wolfe K.H. 2006. Functional partitioning of yeast co-expression networks after genome duplication. PLoS Biol. 4: e109.[CrossRef][Medline] Force A., Lynch M., Pickett F.B., Amores A., Yan Y.L., Postlethwait J. 1999. Preservation of duplicate genes by complementary, degenerative mutations. Genetics 151: 1531–1545. Freeling M. and Thomas B.C. 2006. Gene-balanced duplications, like tetraploidy, provide predictable drive to increase morphological complexity. Genome Res (this issue). Gao L.Z. and Innan H. 2004. Very low gene duplication rate in the yeast genome. Science 306: 1367–1370. Glazko G.V., Koonin E.V., Rogozin I.B., Shabalina S.A. 2003. A significant fraction of conserved noncoding DNA in human and mouse consists of predicted matrix attachment regions. Trends Genet. 19: 119–124.[CrossRef][Medline] Goldberg A.L. 2003. Protein degradation and protection against misfolded or damaged proteins. Nature 426: 895–899.[CrossRef][Medline] Gu Z., Nicolae D., Lu H.H., Li W.H. 2002. Rapid divergence in expression between duplicate genes inferred from microarray data. Trends Genet. 18: 609–613.[CrossRef][Medline] Gu Z., Rifkin S.A., White K.P., Li W.H. 2004. Duplicate genes increase gene expression diversity within and between species. Nat. Genet. 36: 577–579.[CrossRef][Medline] Haberer G., Hindemitt T., Meyers B.C., Mayer K.F. 2004. Transcriptional similarities, dissimilarities, and conservation of cis-elements in duplicated genes of Arabidopsis.. Plant Physiol. 136: 3009–3022. Haldane J.B.S. 1933. The part played by recurrent mutation in evolution. Am. Nat. 67: 5–19.[CrossRef] He X. and Zhang J. 2005. Rapid subfunctionalization accompanied by prolonged and substantial neofunctionalization in duplicate gene evolution. Genetics 169: 1157–1164. Heslop-Harrison J.S. 1990. Gene expression and parental dominance in hybrid plants. Dev. Suppl. 1990: 21–28. Inada D.C., Bashir A., Lee C., Thomas B.C., Ko C., Goff S.A., Freeling M. 2003. Conserved noncoding sequences in the grasses. Genome Res. 13: 2030–2041. Kashkush K., Feldman M., Levy A.A. 2002. Gene loss, silencing and activation in a newly synthesized wheat allotetraploid. Genetics 160: 1651–1659. Kellis M., Birren B.W., Lander E.S. 2004. Proof and evolutionary analysis of ancient genome duplication in the yeast Saccharomyces cerevisiae.. Nature 428: 617–624.[CrossRef][Medline] Koonin E.V. 2005. Orthologs, paralogs, and evolutionary genomics. Annu. Rev. Genet. 39: 309–338.[CrossRef][Medline] Lee H.-S. and Chen Z.J. 2001. Protein-coding genes are epigenetically regulated in Arabidopsis polyploids. Proc. Natl. Acad. Sci. 98: 6753–6758. Lewis E.B. 1951. Pseudoallelism and gene evolution. Cold Spring Harb. Symp. Quant. Biol. 16: 159–174.[Medline] Li W.-H. In Molecular evolution. . 1997. Sinauer, Sunderland, MA. Lippman Z., Gendrel A.-V., Black M., Baughn M.W., Dedhia N., McCombe W.R., Levine K., Mittel V., May B., Kasschau K.D.et al. 2004. Role of transposable elements in heterchromatin and epigenetic control. Nature 430: 471–476.[CrossRef][Medline] Lockton S. and Gaut B.S. 2005. Plant conserved non-coding sequences and paralogue evolution. Trends Genet. 21: 60–65.[CrossRef][Medline] Lynch M. and Force A. 2000. The probability of duplicate gene preservation by subfunctionalization. Genetics 154: 459–473. Maere S., De Bodt S., Raes J., Casneuf T., Van Montagu M., Kuiper M., de Van Peer Y. 2005. Modeling gene and genome duplications in eukaryotes. Proc. Natl. Acad. Sci. 102: 5454–5459. Makova K.D. and Li W.-H. 2003. Divergence in the spatial pattern of gene expression between human duplicate genes. Genome Res. 13: 1638–1645. McLysaght A., Hokamp K., Wolfe K.H. 2002. Extensive genomic duplication during early chordate evolution. Nat. Genet. 31: 200–204.[CrossRef][Medline] Meinke D.W., Meinke L.K., Showalter T.C., Schissel A.M., Mueller L.A., Tzafrir I. 2003. A sequence-based map of Arabidopsis genes with mutant phenotypes. Plant Physiol. 131: 409–418. Ohno S. In Evolution by gene duplication. . 1970. Springer-Verlag, Berlin. Osborn T.C., Pires J.C., Birchler J.A., Auger D.L., Chen Z.J., Lee H.S., Comai L., Madlung A., Doerge R.W., Colot V.et al. 2003. Understanding mechanisms of novel gene expression in polyploids. Trends Genet. 19: 141–147.[CrossRef][Medline] Papp B., Pal C., Hurst L.D. 2003. Dosage sensitivity and the evolution of gene families in yeast. Nature 424: 194–197.[CrossRef][Medline] Paterson A.H., Bowers J.E., de Van Peer Y., Vandepoele K. 2006. Ancient duplications of cereal genomes. New Phytol. 165: 658–661. Raes J. and de Van Peer Y. 2003. Gene duplication, the evolution of novel gene functions, and detecting functional divergence of duplicates in silico. Appl. Bioinformatics 2: 91–101.[Medline] Scannell D.R., Byrne K.P., Gordon J.L., Wong S., Wolfe K.H. 2006. Multiple rounds of speciation associated with reciprocal gene loss in polyploid yeasts. Nature 440: 341–345.[CrossRef][Medline] Seoighe C. and Gehring C. 2004. Genome duplication led to highly selective expansion of the Arabidopsis thaliana proteome. Trends Genet. 20: 461–464.[CrossRef][Medline] Shaked H., Kashkush K., Ozkan H., Feldman M., Levy A.A. 2001. Sequence elimination and cytosine methylation are rapid and reproducible responses of the genome to wide hybridization and allopolyploidy in wheat. Plant Cell 13: 1749–1759. Sokal R.R. and Rohlf F.J. In Biometry. . 1995. W.H. Freeman and Co., New York. Song K., Lu P., Tang K., Osborn T.C. 1995. Rapid genome change in synthetic polyploids of Brassica and its implications for polyploid evolution. Proc. Natl. Acad. Sci. 92: 7719–7723. Stansfield W.D. In The science of evolution. . 1977. Collier Macmillan, New York. Tatusova T.A. and Madden T.L. 1999. BLAST 2 sequences, a new tool for comparing protein and nucleotide sequences. FEMS Microbiol. Lett. 174: 247–250.[CrossRef]< | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
ABSTRACT
Methods
