Synteny conservation between the Prunus genome and both the present and ancestral Arabidopsis genomes

Background Due to the lack of availability of large genomic sequences for peach or other Prunus species, the degree of synteny conservation between the Prunus species and Arabidopsis has not been systematically assessed. Using the recently available peach EST sequences that are anchored to Prunus genetic maps and to peach physical map, we analyzed the extent of conserved synteny between the Prunus and the Arabidopsis genomes. The reconstructed pseudo-ancestral Arabidopsis genome, existed prior to the proposed recent polyploidy event, was also utilized in our analysis to further elucidate the evolutionary relationship. Results We analyzed the synteny conservation between the Prunus and the Arabidopsis genomes by comparing 475 peach ESTs that are anchored to Prunus genetic maps and their Arabidopsis homologs detected by sequence similarity. Microsyntenic regions were detected between all five Arabidopsis chromosomes and seven of the eight linkage groups of the Prunus reference map. An additional 1097 peach ESTs that are anchored to 431 BAC contigs of the peach physical map and their Arabidopsis homologs were also analyzed. Microsyntenic regions were detected in 77 BAC contigs. The syntenic regions from both data sets were short and contained only a couple of conserved gene pairs. The synteny between peach and Arabidopsis was fragmentary; all the Prunus linkage groups containing syntenic regions matched to more than two different Arabidopsis chromosomes, and most BAC contigs with multiple conserved syntenic regions corresponded to multiple Arabidopsis chromosomes. Using the same peach EST datasets and their Arabidopsis homologs, we also detected conserved syntenic regions in the pseudo-ancestral Arabidopsis genome. In many cases, the gene order and content of peach regions was more conserved in the ancestral genome than in the present Arabidopsis region. Statistical significance of each syntenic group was calculated using simulated Arabidopsis genome. Conclusion We report here the result of the first extensive analysis of the conserved microsynteny using DNA sequences across the Prunus genome and their Arabidopsis homologs. Our study also illustrates that both the ancestral and present Arabidopsis genomes can provide a useful resource for marker saturation and candidate gene search, as well as elucidating evolutionary relationships between species.


Background
The eukaryote genome size is vastly diverse and is not dependent on the genetic and organismal complexity. Most of the DNA in large genomes, however, is non-coding and the gene content is relatively constant [1,2]. Arabidopsis thaliana (estimated haploid size of 115 Mb) contains more than 25,000 genes [3], and the Human genome (estimated haploid size of 3200 Mb) contains 20,000-25,000 genes [4]. In addition to the gene content, the conservation in the synteny (the presence of two or more genes in the same chromosome) and gene order has been observed among many plant species. One of the earliest observations of conserved macrosynteny was between potato and tomato in Solanaceae, where cDNA markers along the 12 chromosomes were largely collinear [5].
Significant conservation in the marker and gene order has been observed among grass species, despite the diverse genome size and chromosome numbers [6][7][8]. Similar conserved macrosynteny has also been observed in Rosaceae. Comparisons of anchor markers of the Prunus reference map with those of 13 maps constructed with other Prunus populations showed that the genomes of seven Prunus diploid species are essentially collinear [9]. Large collinear blocks were also detected among different genera in Rosaceae, such as Prunus and Malus [9].
On the other hand, genome sequence comparisons have revealed that plant genome evolution involved various small chromosomal rearrangements, such as insertions, deletions, inversions and translocations [10]. For example, Kilian and coworkers have shown that a barley gene in regions of high microsynteny with rice is in fact transposed to a position that is no longer syntenous with rice [11]. In addition to small chromosomal rearrangements, large segmental duplications and polyploidy is prevalent in plant genome evolution [12][13][14]. Genome duplication was well observed in Brassicaceae; The Brassica genome is extensively triplicated [15] and the Arabidopsis genome contains numerous large duplicated chromosomal segments [3,16]. Comparative physical mapping between Brassica species and Arabidopsis showed high conservation in the gene order but not the gene content, possibly resulting from random gene loss after extensive genome duplication in both genomes [14].
The degree of synteny conservation has also been examined between Arabidopsis and less closely related species. Rosid I and rosid II comparisons ( Figure 1) have been made by sequence homology between soybean marker sequences and Arabidopsis sequences [17]. Shared linkages were identified along with signs of extensive genome duplication and reorganization. A few microsyntenic regions were also identified by comparative physical mapping between Arabidopsis and soybean [18]. A gene-containing BAC sequence of tomato (asteroid I) had conserved synteny with four different segments of Arabidopsis chromosomes 2-5 [19].
Synteny between Arabidopsis and four dicotyledonous species from three major families, caryophyllids, rosids and asteroids, has also been explored by constructing genetic maps based on ESTs that are homologous to Arabidopsis genes [20]. Some syntenic blocks were conserved in all five maps, Arabidopsis, sugar beet, potato, sunflower and Prunus, suggesting their evolutionary significance. The syntenic blocks usually contained only several loci, however, and each linkage group of the crop genetic maps matched to multiple Arabidopsis genome regions. Com-A dendrogram depicting the phylogenetic relationship of peach, Arabidopsis and many other crop species plex syntenic relationships, suggestive of chromosome rearrangement, selective gene loss and genome duplication, were also observed [20]. Synteny between rice and Arabidopsis genomes, after 200 million years of divergence [21], were also observed, but the syntenic regions were scarce and separated by intervening proteins as previously suggested [20]. Also, most of the rice syntenic regions map to more than one Arabidopsis chromosome [21], supporting the theme of large scale genome duplication and selective gene loss in plant genome evolution.
A recent study has systematically analyzed the timing and number of segmental duplications in the Arabidopsis genome and suggested a recent polyploidy superimposed on older large-scale duplication [22]. The recent polyploidy appeared to have occurred during the early emergence of the Brassicaceae family and the older set of duplicated blocks between rosid I and rosid II groups. One of the interesting outcomes from this study is the reconstruction of the approximate gene order of the ancestral genome that existed prior to the recent polyploidy event. The reconstruction was done by merging genes in both sister regions duplicated at the time of polyploidy.
Rosaceae contains numerous important fruit crops such as peach, apple, cherry, pear, raspberry, blackberry and strawberry [23]. Due to the lack of availability of large genomic sequences for peach or other Rosaceae species, little information has been available to study the degree of synteny conservation between the Rosaceae species and Arabidopsis. A recent study has detected fragmentary macrosynteny between the Prunus general map and Arabidopsis, from comparisons of the genetic marker sequences Number of syntenic groups in each TxE linkage group that match to each Arabidopsis chromosome Figure 2 Number of syntenic groups in each TxE linkage group that match to each Arabidopsis chromosome.  and their Arabidopsis homologs [9]. When sequences of three peach genomic regions were used, only short (two or three genes) blocks that are collinear with the Arabidopsis genome were found [24]. With the international effort to make peach the reference species for the Rosaceae family, peach physical mapping is underway and peach ESTs are being anchored to both the genetic and physical map [25].
The objective of this study was to assess the degree of conserved synteny between Prunus and Arabidopsis using these extensive EST sequences anchored to the genetic and physical maps. We also used the reconstructed ancestral Arabidopsis genome to see if we coulc find additional syntenic regions. This study demonstrates that comparative genome analyses between the reconstructed Arabidopsis genome and other plant species can further facilitate the utilization of the genetic resources of both species and help us to understand the evolutionary relationship between these species.

Conserved synteny between Prunus and Arabidopsis
We searched for conserved syntenic regions between the Prunus maps and the Arabidopsis genome using 475 peach ESTs anchored to the Prunus maps and their Arabidopsis homologs detected by a FASTX sequence similarity search (E value less than 10 -5 ). The syntenic groups were selected when the distance between the two adjacent matches were less than 250 kb in the Arabidopsis genome and less than 10 cM in the Prunus maps. We detected 139 conserved syntenic regions, and 20 of them had three or more gene pairs. The number of syntenic regions between Arabidopsis and each of the Prunus maps are shown in Table 1.
Microsyntenic regions were detected between all five Arabidopsis chromosomes and seven of the eight linkage groups of the Prunus TxE reference map. All of the TxE linkage groups which contained syntenic regions matched to more than two different Arabidopsis chromosomes (Figure 2). The gene pairs in the syntenic regions showed significant sequence similarity; 78% had E values less than 10 -15 , and 88% had E values less than 10 -10 .
There were 20 conserved syntenic regions with three or more gene pairs between the Prunus TxE map and the Arabidopsis genome ( Figure 3). Table 2 lists these syntenic regions with the putative functions of the Arabidopsis genes. The largest block (group gp128) had four gene pairs, and covered 20 cM in G2 of the TxE Prunus map and 342 Kb in chromosome 5 of Arabidopsis ( Figure 3). Among 20 regions with three or more gene pairs, five groups showed conserved gene order. In two groups, the collinearity could not be assessed because two different peach ESTs were anchored to the same BAC, probably by hybridizing to different gene sequences in the same BAC. In the rest of the syntenic groups, the gene order was not conserved, suggesting many chromosomal rearrangement events.
Reflecting the synteny conservation among Prunus maps, we detected many Arabidopsis regions matching to more than one Prunus map region. In groups gp42 and gp54, the Arabidopsis genes matched to the ESTs that were anchored to the same markers present in the linkage group G7 of both the TxE Prunus map and the JxF peach map ( Table 2). In groups gp85 and gp98, the Arabidopsis genes within 350 kb matched to ESTs anchored to G2F of the FxT almond map and G5 of the TxE Prunus map ( Table  2).
Most of the peach ESTs showed strong similarity to more than one Arabidopsis genes, and we were able to detect Prunus blocks that map to more than one site in the Arabidopsis genome. Interestingly, some of these putative duplicated Arabidopsis regions were located in the Arabi-Conserved syntenic regions with three or more gene pairs between Arabidopsis genome and Prunus genome Figure 3 Conserved syntenic regions with three or more gene pairs between Arabidopsis genome and Prunus genome. Bolded blocks are the ones with conserved gene order.
dopsis paralogous blocks -duplicated blocks in a genome -reported in the previous study [21]. Figure 4 shows those Prunus blocks, syntenic to two different Arabidopsis regions, juxtaposed to the plot of the paralogous blocks of Arabidopsis. All three paralogons were the ones that were generated by a recent polyploidy event that occurred during the early emergence of the Brassicaceae. Arabidopsis blocks with conserved synteny to a region in FxT-G1F and JxF-G1 belong to the paralogons in chromosome 1 and 4, and those with conserved synteny to a region in FxT-G2T belong to the paralogons in two different arms of chromosome 5 ( Figure 4). Three distinct regions in TxE -linkage groups G2, G4 and G5 -showed conserved synteny to three overlapping blocks in each paralogon on chromosome 4 and 5 ( Figure 4). These TxE map regions may rep-resent triplicated Prunus regions that subsequently went through selective gene loss.

Synteny between Prunus and the pseudo ancestral Arabidopsis genome
To further analyze the evolutionary relationship between the Arabidopsis and Prunus genomes, we searched for conserved syntenic regions between Prunus maps and the ancestral Arabidopsis genome [22]. The pseudo ancestral genome contained 20187 genes, which is about 69% of the genes in the present genome, arranged in a linear array. We used the same 475 peach ESTs and their Arabidopsis homologs detected by FASTX sequence similarity searching (E value less than 10 -5 ) in our search for the conserved syntenic regions. The syntenic groups were selected when the number of genes between the two adja-Prunus genomic blocks that map to two distinct Arabidopsis regions Figure 4 Prunus genomic blocks that map to two distinct Arabidopsis regions. Shown are the Prunus blocks that identified Arabidopsis sister regions generated by the proposed polyploidy event. The Prunus blocks with the same color (red or green) are homologous regions that share more than two anchored ESTs.
cent matches is less than 61 in the Arabidopsis genome and the distance less than 10 cM of the Prunus maps. The estimated number of genes in 250 kb was used as the maximum distance between two matches in the Arabidopsis genome, since only the gene order, instead of the kb, was available as a position along the ancestral genome (see Methods).
We detected 101 conserved syntenic regions, and 12 of them had three or more gene pairs. The details, including the putative functions of the syntenic blocks with three or more gene pairs, are shown in Table 3. Fewer syntenic blocks were detected in the ancestral genome using these criteria, but much fewer blocks matched to the duplicated Arabidopsis genome. In the present Arabidopsis genome, 20 syntenic blocks, with three conserved genes, matched to 14 distinct Prunus regions, but, in the ancestral genome, 12 syntenic blocks matched to 10 distinct Prunus regions. Some groups contained the same Arabidopsis gene and peach EST pairs as in the syntenic groups detected from the Prunus-present Arabidopsis genome analysis. Several new Prunus regions were found to have conserved synteny with the ancestral Arabidopsis genome. The Arabidopsis genes in these syntenic blocks were apparently relocated in distinct regions after the putative Arabidopsis genome duplication event. For example, group ga54 in ancestral genome is composed of two genes in chromosome 5 and one from chromosome 3, and they were paired with ESTs that were anchored to the linkage group G1 of TxE map. Group ga28 and ga79 represent regions where three genes were closely located in the ancestral genome but they were rearranged into two different regions of the present Arabidopsis chromosome 5.
We also found examples where the gene content in the Prunus genome is more conserved in the ancestral genome than the present Arabidopsis genome. For example, group ga81 in ancestral genome contains four gene pairs that match to the linkage group G5 of the TxE map ( Figure 5). Group gp48 and gp101 in the present genome match to the same region in TxE-G5, but contain only part of the gene pairs. Figure 5 illustrates the proposed evolutionary steps that may have occurred in these regions: large scale genome duplication and subsequent selective gene loss and gene duplication. The genomic regions in chromosome 2 and 4 were part of the previously reported duplicated regions with 68 gene pairs [22], supporting our proposed evolutionary steps.

Synteny analysis between the peach physical transcriptome map and the Arabidopsis genome
We also used peach EST sequences that are anchored to the developing peach physical map to search for conserved syntenic regions between peach and Arabidopsis.
Our data were composed of 1097 peach ESTs that are anchored to 431 BAC contigs, and their Arabidopsis homologs detected by FASTX sequence similarity searching (E value less than 10 -5 ). The sequence similarity search results produced 4448 peach-Arabidopsis sequence pairs that consist of 904 distinct ESTs and 3747 distinct Arabidopsis proteins. These sequence pairs were used to detect syntenic regions between peach and Arabidopsis. The syntenic groups were selected when the distance between the two adjacent matches was less than 250 kb in the Arabidopsis genome and anchored to the same BAC contig.
Our analysis identified 287 Arabidopsis genes and 204 peach ESTs found in 140 syntenic blocks with at least two gene pairs. The syntenic blocks were found in all of the five Arabidopsis chromosomes. In peach, the syntenic blocks were found in a total of 77 BAC contigs. The synteny conservation was fragmentary; 16 out of the 18 BAC contigs with multiple syntenic regions matched to more than one Arabidopsis chromosome.
The number of gene pairs in the syntenic blocks was small: two blocks with four gene pairs, 14 blocks with three gene pairs and 124 blocks with two gene pairs. The syntenic blocks with three or more gene pairs are shown in Table 4 and Figure 6. Only two of the 16 blocks were collinear. It is possible that the content in the block is conserved but the gene order has differentially evolved in the two genomes. On the other hand, the order of the peach ESTs was estimated by the positions of the EST-hybridizing BACs in a BAC contig which may not represent the actual order of the ESTs in the genome. The average size of the syntenic blocks in Arabidopsis genome was 97 kb with a maximum 360 kb (group pp96: Arabidopsis chromosome 4 and ctg2264) and minimum 2.7 kb. Groups pp129 and pp130 were close enough to be combined into one syntenic region containing five gene pairs, and they covered 451 kb in the Arabidopsis genome ( Figure 6).  the rest of the ESTs may represent genes that do not exist in the Arabidopsis gene repertoire. However, it is also possible that we will detect more Arabidopsis homologs, hence more microsyntenic regions, when the entire gene sequences are available instead of short EST sequences.
In addition to the blocks in ctg2264, we found many other peach blocks corresponding to more than one syntenic region in Arabidopsis, reflecting the fact that the Arabidopsis genome contains numerous large duplicated segments [21]. In our data set, there were 21 peach segments that each corresponds to more than one distinct Arabidopsis segment. As expected, the Arabidopsis genes that matched to the same peach ESTs in these duplicated regions had similar putative function or belong to the same protein family. Some of the syntenic blocks, especially those duplicated in the Arabidopsis genome, were composed of genes with related function, suggesting that related genes that tend to cluster in Arabidopsis also do in peach. For example, all four Arabidopsis genes in groups pp77 and pp110 were FAD-binding domain-containing protein, similar to reticuline oxidase precursor. Similar observation has been reported in the analysis between Arabidopsis and rice [25]. We also observed two Arabidopsis segments that each corresponds to more than one distinct peach segment. Groups pp113 and pp132 involve an Ara-bidopsis region with three genes in chromosome 5 matching three peach ESTs in two different contigs (ctg1505 and ctg2269) and groups pp114 and pp123 involve an Arabidopsis region that matches to two different peach contigs (ctg1565 and ctg2287).

Synteny analysis between the peach physical transcriptome map and the reconstructed Arabidopsis ancestral genome
The evolutionary relationship between Arabidopsis and peach was further analyzed by searching for conserved syntenic regions between the ancestral Arabidopsis genome and the peach physical transcriptome map. The syntenic groups were selected when the number of genes between the two adjacent matches was less than 61 in the Arabidopsis genome and anchored to the same BAC contig. This analysis identified 231 Arabidopsis proteins and 179 peach ESTs found in 111 conserved gene blocks. The average block size in the Arabidopsis genome was 27.6 genes with a maximum of 97 genes and a minimum of two genes. The estimated size of the syntenic blocks, using the average size of the Arabidopsis genome containing one gene per 4.1 kb (see Methods), is on average 113.2 kb with a maximum 397.7 kb and a minimum of 8.2 kb. The syntenic blocks were distributed quite evenly across the ancestral genome. In peach, the syntenic blocks were found in a Proposed evolutionary steps involving some syntenic blocks between Arabidopsis and the Prunus genomes Figure 5 Proposed evolutionary steps involving some syntenic blocks between Arabidopsis and the Prunus genomes. Blocks in the putative ancestral Arabidopsis genome and Arabidopsis chromosome 2 and 4 that match to the same block in Prunus TxE map are illustrated. Red and green colors were used to help track the genes. Dashed lines were used to indicate the relationship with less stronger homology when the same EST was homologous to more than one Arabidopsis genes. total of 69 contigs. Among the 111 syntenic blocks, two blocks had four gene pairs, 12 blocks had three gene pairs and the rest had two gene pairs. The details of the 12 blocks with three or more gene pairs are shown in Table  5. Four of the 12 blocks with three or more gene pairs were collinear. Five groups contained the same Arabidopsis gene and peach EST pairs as those in the syntenic groups detected from the peach-present Arabidopsis genome analysis. Four groups involved the same regions to the ones observed in the peach-present Arabidopsis genome analysis, except that one or two peach ESTs were paired with Arabidopsis proteins from other duplicated regions. The rest of the blocks disclose peach regions that have conserved synteny with the ancestral Arabidopsis genome but not with the present one. In group pa3, AT5G60910 and the other two genes are closer in the ancestral genome, with only four genes in between, than in the present genome where they are 21 Mbp apart from each other. Groups pa5 and pa35 shows a similar situation in which three genes are far apart in the same chromosome of the present genome, but they are much closer in the ancestral genome.
Ctg2264, containing the most anchored ESTs, had one with four unordered gene pairs, four with three unordered gene pairs and 18 with two gene pairs. Upon close examination, the syntenic block with the five unordered genes observed in the present Arabidopsis genome ( Figure 6) was also detected in the ancestral genome (Figure 7). The block was not detected from our original analysis because some of the gaps between the genes were larger than the limit set by the search parameters. The comparison revealed a syntenic block with six gene pairs in the ancestral genome and two blocks containing rearranged gene pairs in chromosome 3 and 5 of the present Arabidopsis genome (Figure 7). Figure 7 illustrates the proposed evolutionary steps that may have occurred in these regions: large scale genome duplication and subsequent selective gene loss in chromosome 3 and inversion in chromosome 5. Since the reconstructed ancestral Arabidopsis genome has been reported to contain a considerable amount of duplicated regions [22], we searched for peach EST segments that paired with more than one distinct Arabidopsis region. In this data set, there were eleven peach segments that each corresponds to two distinct Arabidopsis seg-Conserved syntenic regions with three or more gene pairs between Arabidopsis genome and EST-anchored peach BAC contigs Figure 6 Conserved syntenic regions with three or more gene pairs between Arabidopsis genome and EST-anchored peach BAC contigs.
ments. It is notable, however, that twice as many duplicated blocks were identified by the peach EST segments in the present genome than the ancestral genome. We also observed three Arabidopsis segments that each corresponded to more than one distinct peach segment. Two Arabidopsis segments identified the same duplicated peach segments, detected from the analysis with the present Arabidopsis genome. Another Arabidopsis region identified duplicated peach regions in ctg1112 and ctg2175.

Simulation study
To determine whether the syntenic groups we report were detected by chance, we tested the statistical significance for each group. Both the current and putative ancestral Arabidopsis genomes were randomized by leaving the locations the same but permuting the gene names. We analyzed 1000 simulated Arabidopsis genomes for the occurrence of the each conserved syntenic group and calculated the probability of the match occurring by chance. The probability of the association by chance was less than 1% for all the syntenic groups with more than three gene pairs. The numbers of syntenic groups at various significance thresholds are shown in Table 6.

Discussion
We surveyed the degree of synteny conservation between the Prunus and the Arabidopsis genomes using extensive EST sequences anchored to several Prunus genetic maps and the developing peach physical map. Our study is the first to systematically examine the conserved microsynteny using DNA sequences across the Prunus genome and their Arabidopsis homologs. We could detect considerable conserved microsytenic regions even with our stringent parameters. Among the 475 genetically anchored ESTs, 142 distinct ESTs belong to the syntenic groups that were conserved with either the present or ancestral Arabidopsis genomes. However, the syntenic blocks were rather small in size and contained only a few gene pairs. In addition, most of the BAC contigs with more than two conserved syntenic regions matched to more than one Arabidopsis chromosome. Our finding is in accordance with the previous study of peach BAC sequences that the segments with a gene order congruent with Arabidopsis were short in any peach region studied and the corresponding segments were found in diverse locations in the Arabidopsis genome [24]. From the analysis with the genetically anchored ESTs, the largest block we detected had four gene pairs, and covered 20 cM in G2 of the TxE Prunus map and 342 Kb in chromosome 5 of Arabidopsis. From the analysis with the physical map-anchored ESTs, the largest block we detected contained five gene pairs and spanned 451 kb in the Arabidopsis genome. We may be able to find more syntenic blocks with over three gene pairs when more ESTs are hybridized to map-anchored BACs and longer BAC contigs are available. We may also find more syntenic blocks when the entire gene sequences are available. The results from the BAC contig rich in anchored ESTs, however, suggest that the syntenic regions between Arabidopsis and peach are typically small and contain several gene pairs at most. For example, ctg2264, with five BACs and 70 anchored ESTs, have numerous microsyntenic regions in all five Arabidopsis chromosomes instead of having relatively large syntenic regions.
We also detected conserved syntenic regions in the pseudo ancestral Arabidopsis genome that existed prior to the Proposed evolutionary steps involving some syntenic blocks between Arabidopsis and Peach genomes Figure 7 Proposed evolutionary steps involving some syntenic blocks between Arabidopsis and Peach genomes. Blocks in the putative ancestral Arabidopsis genome and Arabidopsis chromosome 3 and 5 that match to the same peach BAC contig are illustrated. Red colors were used to help track the genes. The order of the ESTs in the BAC contig was not shown because the ESTs were anchored to overlapping BACs. recent polyploidy event. We did not find markedly different results in the conserved synteny with the ancestral genome compared to the present genome, which was to be expected given that the polyploidization event that differentiated the present and the ancestral Arabidopsis genome occurred 24-40 million years ago, which is relatively recent compared to the peach-Arabidopsis divergence, 90 million years ago. We did find, however, a number of syntenic regions in the ancestral genome that do not exist in the present genome. We also found some examples where gene content and the gene order is more conserved in the ancestral genome than in the present genome. Our study illustrates that comparative genome analysis of both the ancestral and present Arabidopsis genomes with other plant species can provide a useful resource for marker saturation in a specific region and candidate gene searches, as well as elucidating evolutionary relationships between species.

Conclusion
We report the results of the systematic examination of conserved microsynteny between the Prunus and Arabidopsis. Our study is the first to systematically examine the conserved microsynteny using extensive DNA sequences across the Prunus genome and their Arabidopsis homologs. More importantly, this study utilized the pseudo-ancestral Arabidopsis genome, as well as the present Arabidopsis genome, in the comparison of the Arabidopsis with other plant genomes. This method helped us to find more conserved microsyntenic regions between the ancestral Arabidopsis and Prunus genomes and also to delineate the putative evolutionary steps in the microsyntenic regions. We believe that this report will give a new insight in the study of evolutionary relationships among plants and provide new way to more efficient utilization of the resources of the model genome.

Data description
For the synteny analysis between the Prunus and Arabidopsis genomes, we used peach EST sequences anchored to the Prunus genetic maps [25]. Among the 475 genetically anchored peach ESTs used in this analysis, 306 ESTs were hybridized to BACs that have been hybridized to genetic markers, and the rest were hybridized to BACs belonging to a contig containing other BACs hybridized to genetic markers. The positions (cM) of the genetic markers were used as the positions for the genetically anchored ESTs.
For the synteny analysis between the peach physical transcriptome map and Arabidopsis, we used peach EST sequences that are anchored the developing peach physical map. The data set is composed of 1097 sequences that are anchored to 431 BAC contigs containing at least two anchored ESTs. The position of the individual BACs in the BAC contigs were used as the positions of the physical map anchored ESTs. For the ESTs that are anchored to multiple overlapping ESTs in a BAC contig, the innermost left and right positions were assigned. All the sequences and positions of the peach ESTs were obtained from the Genome Database for Rosaceae (GDR) [27,28].
The sequence data (ATH1_pep_cm_20040228) and the chromosome coordinate data (sv_gene.data) of the 29161 Arabidopsis translated proteins were downloaded from the Arabidopsis Information Resources (TAIR) database [29,30] in March 2005. The ordered list of 20187 gene names in the reconstructed ancestral Arabidopsis genome was downloaded from the Paralogons in Arabidopsis thaliana web site [22,31].

Detection of the conserved syntenic regions
Mapped peach ESTs that are homologous to the Arabidopsis proteins were determined using the FASTX 3.4 algorithm [27]. Matches with E values less than 10 -5 were selected for further analysis. For the comparison between the Arabidopsis genome and the Prunus maps, the syntenic groups were selected when the distance between the two adjacent matches were less than 250 kb in the Arabidopsis genome and less than 10 cM for the Prunus maps. For the comparison between the Arabidopsis genome and the peach physical map, the syntenic groups were selected when the matches were located within 250 kb in the current Arabidopsis genome and belong to the same BAC contigs. In the analysis of the conserved synteny between the ancestral Arabidopsis genome and the peach physical map  (14) Numbers in parenthesis stands for the syntenic groups with more than three gene pairs.
or the Prunus genetic maps, we used the estimated number of genes in 250 kb (61 genes) as the maximum distance between the two adjacent matches in the Arabidopsis genome. The estimation was done by dividing 250 kb by the average size per gene (4.1 kb) in Arabidopsis, which is derived by the division of the total length in kb by the number of genes in the Arabidopsis genome.
We used a program called DAGchainer [32] to detect collinear chromosomal segment conserved in the peach/Prunus and Arabidopsis genomes. DAGchainer was run with parameters set to detect any collinear blocks with two or more gene pairs and with the maximum distance between the two adjacent matches specified above. Since the DAGchainer program detects only the regions with conserved order, we developed scripts to detect both collinear and non-collinear regions from the output.

Evaluation of the conserved syntenic regions
To determine whether the syntenic groups we report were detected by chance, we tested the statistical significance for each group. Both of the current and putative ancestral Arabidopsis genomes were randomized by leaving the locations the same but permuting the gene names. We analyzed 1000 simulated Arabidopsis genomes for the occurrence of each conserved syntenic group and calculated the probability of the match occurring by chance.

Authors' contributions
SJ designed the protocol for synteny analysis and the statistical analysis, designed and developed scripts, performed the research, analyzed the data and wrote the paper. DM conceived of the study and participated in its design and coordination, and critically revised the manuscript. MS performed the sequence similarity search and wrote the scripts for statistical analysis. IC wrote the scripts for detecting non-linear syntenic regions and duplicate syntenic regions and parting the DAGchainer outputs. TZ provided the EST data hybridized to peach BAC contigs. PA critically revised the manuscript. AA conceived of the study and critically revised the manuscript. All authors read and approved the final manuscript.
Publish with Bio Med Central and every scientist can read your work free of charge