Construction, alignment and analysis of twelve framework physical maps that represent the ten genome types of the genus <it>Oryza</it>

gb-2008-9-2-r45 GBJ Method Construction, alignment and analysis of twelve framework physical maps that represent the ten genome types of the genus <it>Oryza</it> Kim HyeRan kimhr@cnu.ac.kr Hurwitz Bonnie bonnie.hurwitz@gmail.com Yu Yeisoo yeisooyu@ag.arizona.edu Collura Kristi kcollur@ag.arizona.edu Gill Navdeep gilln@purdue.edu SanMiguel Phillip pmiguel@purdue.edu Mullikin C James mullikin@mail.nih.gov Maher Christopher camaher@gmail.com Nelson William will@agcol.arizona.edu Wissotski Marina marinaw@ag.arizona.edu Braidotti Michele michele.braidotto@gmail.com Kudrna David dkudrna@ag.arizona.edu Goicoechea Luis José jlgoicoe@ag.arizona.edu Stein Lincoln lstein@cshl.org Ware Doreen ware@cshl.edu Jackson A Scott sjackson@purdue.edu Soderlund Carol cari@agcol.arizona.edu Wing A Rod rwing@ag.arizona.edu

Arizona Genomics Institute, Department of Plant Sciences, University of Arizona, Tucson, Arizona 85721, USA

Cold Spring Harbor Laboratory, Cold Spring Harbor, New York 11724, USA

Department of Agronomy, Purdue University, West Lafayette, Indiana 47907, USA

Department of Horticulture, Purdue University, West Lafayette, Indiana 47907, USA

Genome Technology Branch, NHGRI, National Institutes of Health, Bethesda, Maryland, 20892, USA

Department of Biomedical Engineering, Stony Brook University, Stony Brook, New York 11794, USA

Arizona Genomics Computational Laboratory, University of Arizona, Tucson, Arizona 85721, USA

USDA-ARS NAA Plant, Soil and Nutrition Laboratory Research Unit, Ithaca, New York 14853, USA

Genome Biology 1465-6906 2008 9 2 R45 http://genomebiology.com/2008/9/2/R45 18304353 10.1186/gb-2008-9-2-r45 14 12 2007 12 2 2008 28 2 2008 28 02 2008 2008 Kim et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Oryza framework physical maps

Bacterial artificial chromosome (BAC) fingerprint and end-sequenced physical maps representing the ten genome types of Oryza are presented

Abstract

We describe the establishment and analysis of a genus-wide comparative framework composed of 12 bacterial artificial chromosome fingerprint and end-sequenced physical maps representing the 10 genome types of Oryza aligned to the O. sativa ssp. japonica reference genome sequence. Over 932 Mb of end sequence was analyzed for repeats, simple sequence repeats, miRNA and single nucleotide variations, providing the most extensive analysis of Oryza sequence to date.

Evolution Genetics Genome studies Plant biology

Background

Comparative genomics is a powerful tool for unraveling the evolutionary history and gene functionality of related species. The availability of high resolution genetic and physical maps and genome assemblies has established comparative genomics platforms in organisms ranging from bacteria and fungi to animals and plants 1234567. The majority of sequence-based eukaryotic platforms have focused on comparisons between genera 891011. Although highly informative, the divergence times between genera are often too distant to be useful for the identification of conserved noncoding sequences such as transcription factor binding sites, enhancers and matrix attachment regions (for example, see 1213). Sequence comparisons between species within a single genus have focused primarily on orthologous loci or genomic regions 1415. Whole genome comparative platforms within a genus are still in their infancy, with the most notable exception being the generation of whole genome shotgun assemblies of 12 Drosophila species that span an evolutionary time frame of approximately 60 million years 16.

In plants, rice (Oryza sativa) and thale cress (Arabidopsis thaliana) are important model systems for functional and evolutionary biology. Both genomes have been completely sequenced 1718 and serve as the reference sequences for the major monocot and dicot plant lineages. Rice has added significance by virtue of being the world's most important crop, directly feeding half the human population. The population that depends on rice is expected to double in 25-50 years and there is thus an intense effort by breeders to double current rice yields using less land and water and on poorer soils 19. To achieve this goal, the plant biology community is engaged in a concerted effort to functionally characterize all plant genes in both model plants using approaches including genetics, transgenetics, and comparative genomics. These efforts have significance not only to world food and bio-energy security issues but also to fundamental eukaryotic biology.

Comparative genomics across the cereals and within the genus Oryza will play a major role in these efforts. Oryza is composed of 24 species, including 2 domesticated (O. sativa, and O. glaberrima) and 22 wild species 2021. These species are classified into ten genome types (six diploids and four allotetraploids) based on crossing barriers 2223, chromosome pairing 2425, morphology 2627, and molecular phylogenetics 28. Compared to the domesticates, the wild relatives of rice are phenotypically inferior grass-like plants, but are a virtually untapped reservoir of agriculturally important genes and allelic variants that can be used to improve cultivated rice 2129. Therefore, comparative genomics of wild and domesticated Oryza species may not only provide useful materials for breeding, but will also shed light on the evolution and domestication of cultivated rice.

Our long-term objective is to develop a genome-level comparative experimental system for the genus Oryza (the Oryza Map Alignment Project (OMAP)). Our strategy is to develop bacterial artificial chromosome (BAC)-end sequence (BES) physical frameworks of the ten genome types of Oryza and align them to the reference sequence. These frameworks can then be populated with an array of phenotypic, genetic, biochemical, and physiological data to address fundamental questions in biology and agriculture. Previously, we described the construction and characterization of 12 BAC libraries from representative species that encompass the 10 genome types of Oryza 30. Here we report the construction of 12 BAC/BES framework physical maps derived from these libraries and an analysis of the BESs in terms of transposable element, simple sequence repeat (SSR), microRNA (miRNA), and single nucleotide variation (SNV) content.

Results

Fingerprinting, BAC end sequencing and contig assembly to generate phase I physical maps for the ten genome types of Oryza

BAC-based physical maps for the ten genome types of Oryza were generated using a two-step procedure. First, clones from 12 Oryza BAC libraries 30 were fingerprinted 31 and BAC end sequenced. Fingerprints from each library were then assembled into contigs using FPC 32. Fingerprint, BES and FPC assembly statistics are summarized in Tables 1 and 2. Briefly, we generated a total of 710,536 fingerprints and 1,452,912 BESs (approximately 932 Mb). Fingerprinting and BES success rates ranged from 85-94% and 88-96%, respectively. BESs represented approximately 8-17% of each Oryza genome, corresponding to one sequence tag per every 4-8 kb. From 90-98% of the BESs were paired and ranged in sequence length from 559 bp in O. minuta [BBCC] to 717 bp in O. officinalis [CC]. More importantly, an average of 89% of the fingerprinted clones had paired BESs. An average of 95% of the fingerprints assembled into contigs, with the remainder existing as singletons. The number of assembled contigs for each species ranged from 422 in O. brachyantha [FF] to 3,962 in O. minuta [BBCC], which roughly correlated with the size of each genome. The 12 unedited FPC assemblies are defined as phase I physical maps (Table 2) and are available interactively using WebFPC 33 or CMap, or as downloadable FPC files (see Materials and methods).

Table 1

Summary of BAC end sequences of 12 Oryza species

Species

Genome type

Genome size (Mb)*

No. of GenBank submissions

Average length after trim (in GenBank)

Total sequenced length (in GenBank)

Genome coverage

No. of forward reads

No. of reverse reads

No. of clones with paired reads (% of total BES)

O. nivara

448

106,124

665 bp

~ 71 Mb

16%

53,450

52,674

51,820 (98%)

O. rufipogon

439

70,982

704 bp

~ 50 Mb

11%

35,473

35,509

34,747 (98%)

O. glaberrima

357

66,821

590 bp

~ 39 Mb

11%

33,456

33,365

30,885 (92%)

O. punctata

425

68,384

710 bp

~ 49 Mb

11%

35,051

33,333

32,284 (94%)

O. officinalis

651

101,091

717 bp

~ 72 Mb

11%

49,972

49,052

47,197 (93%)

O. minuta

BBCC

1,124

169,460

559 bp

~ 95 Mb

85,466

83,994

82,248 (97%)

O. alta

CCDD

1,008

128,732

586 bp

~ 75 Mb

64,365

64,367

58,217 (90%)

O. australiensis

965

135,769

625 bp

~ 85 Mb

67,245

68,524

64,081 (94%)

O. brachyantha

362

67,364

672 bp

~ 45 Mb

13%

34,009

33,355

32,258 (96%)

O. granulata

882

138,171

674 bp

~ 93 Mb

11%

69,962

68,209

66,434 (96%)

O. ridleyi

HHJJ

1,283

204,729

632 bp

~ 129 Mb

10%

102,640

102,089

98,160 (96%)

O. coarctata

HHKK

771^†

195,285

661 bp

~ 129 Mb

17%

100,341

94,944

91,853 (94%)

Total/Average

1,452,912

650 bp

~ 932 Mb

11%

731,430

719,415

690,184 (95%)

*From [30]. ^†By K Arumuganathan using flow cytometric methods as previously described [30].

Table 2

Summary of phase I FPC physical maps of 12 Oryza species

Species

Average insert size

No. of total attempted

No. of clones FPCed

Genome coverage by all FPCed clones

Success ratio

Total organellar contam + no. inserts containing clones*

No. of FPC clones with paired BES reads

No. of singletons (%)

No. of contigs

Total CB units

Average no. of bands/clone

Deduced size of 1 CB unit^†(kb)

Deduced genome size (Mb)^‡(coverage)

O. nivara

161

55,296

51,056

92%

4.1%

48,032 (94%)

2,356 (4.6%)

456

384,733

121.6

1.32

509 (114%)

O. rufipogon

134

35,712

33,023

92%

3.9%

31,202 (94%)

1,305 (4.0%)

637

406,768

106.2

1.26

513 (117%)

O. glaberrima

130

36,864

33,065

90%

3.7%

27,661 (84%)

2,098 (6.3%)

905

309,740

103.9

1.25

388 (109%)

O. punctata

142

36,864

34,224

93%

1.8%

30,228 (88%)

1,482 (4.3%)

490

340,240

117.4

1.21

412 (97%)

O. officinalis

141

54,144

45,856

85%

6.7%

40,549 (88%)

1,133 (2.5%)

703

462,126

115.5

1.22

564 (87%)

O. minuta

125

92,160

86,861

94%

1.2%

77,999 (90%)

9,576 (11.0%)

3,962

1,004,697

96.6

1.29

1,300 (116%)

O. alta

133

73,728

63,860

87%

0.4%

50,842 (80%)

3,111 (4.9%)

2,492

870,411

109.3

1.22

1,059 (105%)

O. australiensis

152

73,728

67,416

91%

2.7%

58,992 (88%)

3,144 (4.7%)

1,003

691,155

125.1

1.22

840 (87%)

O. brachyantha

131

36,864

33,424

91%

1.7%

29,523 (88%)

2,354 (7.0%)

422

215,208

100.9

1.30

279 (77%)

O. granulata

134

73,728

64,836

88%

3.4%

58,818 (91%)

3,032 (4.7%)

2,358

1,078,237

120.8

1.11

1,196 (136%)

O. ridleyi

127

110,592

104,393

94%

2.1%

93,192 (89%)

5,169 (5.0%)

2,190

1,216,461

111.2

1.14

1,389 (108%)

O. coarctata

123

100,224

92,522

92%

2.1%

82,180 (89%)

1,810 (2.0%)

1,250

611,953

100.0

1.23

753 (98%)

Total/Average

136

779,904

710,536

91%

2.8%

629,218 (89%)

5.1%

110.7

1.23

*From [30]. ^†Deduced one CB unit size = average inset size/average number of bands. ^‡Deduced genome size = ^†Deduced one CB unit size × Total CB units.

To determine the approximate genome coverage of each of the 12 phase I physical maps, we estimated the size of the ordered consensus bands for each map in kilobases based on the average insert size of each BAC library and the average number of fingerprint bands observed for each species. Using this estimation, the genome coverage of the phase I physical maps ranged from 136% in O. granulata [GG] to 77% in O. brachyantha [FF]. In total, 7 of the physical maps had coverages greater that 100% (O. nivara, O. rufipogon, O. glaberrima, O. minuta, O. alta, O. granulata, and O. ridleyi), 4 had coverages between 98% and 87% (O. punctata, O. officinalis, O. australiensis, and O. coarctata) and 1 had a coverage of 77% (O. brachyantha). Coverage discrepancies above or below 100% could be due to a number of parameters, including: possible overlaps between the contigs that were not merged by the criteria used in this study; inaccurate estimation of genome size; and over-estimation of average insert sizes of the BAC libraries that would lead to an over-estimation of kilobases per consensus band (CB).

Alignment of phase I physical maps to the O. sativa reference sequence

To evaluate the coverage of the phase I physical maps graphically, we aligned the BESs associated with each FPC contig to the O. sativa reference sequence (RefSeq) 18 using SyMAP 34. SyMAP displays for all 12 physical maps can be viewed at 35 (display details are in Additional data file 1) and are summarized in Table 3 and Additional data file 2. Overall, between 61% and 95% of the contigs from each of the phase I physical maps could be aligned to the O. sativa RefSeq. Contigs not assigned a chromosomal position were generally small (133-289 CB/contig) and had few clones per contig (3-38 clones/contig) (Table 3). The percentage of BESs aligned, under identical parameters, was the least in O. alta [CCDD] (8% alignment) and the greatest in the AA genome type species O. nivara (60% alignment). Interestingly, the FF (O. brachyantha) and HHKK (O. coarctata) genome type species showed substantially more alignments to the O. sativa RefSeq than more closely related species (CC, BBCC, CCDD, EE, GG, and HHJJ).

Table 3

Summary of alignments of 12 OMAP Phase I FPC maps to the O. sativa RefSeq

Not aligned contig

Species

No. of contigs aligned

Total CB aligned

Total size aligned* (Mb)

No. of clones in aligned contig

No. of BESs aligned

Average identity of BES alignments

E-value of BES alignments

CB/contig (average)

No. of clones/contig (average)

O. nivara

368 (81%)

371,561 (97%)

490 (109%)

48,424 (99%)

63,293 (60%)

97%

e-237 to e-33

150

O. rufipogon

581 (91%)

399,346 (98%)

503 (115%)

31,398 (99%)

36,725 (52%)

97%

e-224 to e-57

133

O. glaberrima

860 (95%)

302,696 (98%)

378 (106%)

30,708 (99%)

33,356 (50%)

97%

e-213 to e-42

157

O. punctata

462 (94%)

332,224 (98%)

402 (95%)

31,674 (97%)

25,842 (38%)

93%

e-183 to e-30

286

O. officinalis

617 (88%)

448,477 (97%)

547 (84%)

44,194 (99%)

25,699 (25%)

93%

e-204 to e-16

159

O. minuta

3,073 (78%)

833,710 (83%)

1,075 (96%)

66,277 (86%)

46,227 (27%)

93%

e-173 to e-07

192

O. alta

1,571 (63%)

671,625 (77%)

819 (81%)

50,281 (83%)

10,770 (8%)

93%

e-193 to e-02

216

O. australiensis

787 (78%)

630,836 (91%)

770 (80%)

60,538 (94%)

20,429 (15%)

93%

e-182 to e-03

279

O. brachyantha

351 (83%)

205,608 (96%)

267 (74%)

30,717 (99%)

25,387 (38%)

91%

e-175 to e-18

135

O. granulata

1,444 (61%)

814,383 (76%)

904 (102%)

50,117 (81%)

18,785 (14%)

91%

e-179 to e-05

289

O. ridleyi

1,756 (80%)

1,123,902 (92%)

1,281 (100%)

94,565 (95%)

37,300 (18%)

92%

e-170 to e-06

213

O. coarctata

1,002 (80%)

579,809 (95%)

713 (92%)

89,775 (99%)

53,920 (28%)

92%

e-184 to e-07

130

*Deduced from [1 CB unit size from Table 2 × Total CB unit aligned].

Figure 1 shows a SyMAP alignment of chromosome 1 of the eight diploid Oryza species to the O. sativa RefSeq as an example. As expected, the most co-linear map alignments were observed with species that were evolutionarily less divergent (for example, AA, BB, CC). Regions of disrupted colinearity were detected mostly in the centromeric regions of the O. sativa genome. Although SyMAP alignments were also performed for the four polyploid species, we cannot conclude anything conclusive about their genome organization until their physical maps can be manually subdivided into each of their subgenomes.

Figure 1

SyMAP view of unedited physical maps of chromosome 1 from eight diploid Oryza species aligned to the O. sativa ssp. japonica chromosome 1 RefSeq

SyMAP view of unedited physical maps of chromosome 1 from eight diploid Oryza species aligned to the O. sativa ssp. japonica chromosome 1 RefSeq. The numbers in the small rectangles on the left are contig numbers of OMAP phase I physical maps. Beige bars on the right represent the O. sativa RefSeq (IRGSP V.4 assembly) and the red crosses on the beige bars represent the CentO position of O. sativa [18]. Purple lines represent BESs aligned to the O. sativa RefSeq. The order of the species from the left to right is; O. nivara [AA], O. rufipogon [AA], O. glaberrima [AA], O. punctata [BB]; O. officinalis [CC], O. australiensis [EE]; O. brachyantha [FF], O. granulata [GG].

Sequence analysis of the OMAP BES data set: repeats, SSRs, miRNAs, and SNV

The large BES data set produced by this project offered a unique opportunity to explore sequence content and variation across the Oryza. We next investigated repeat and SSR content across all twelve species and identified miRNA content and SNV amongst the three AA and one BB genomes.

Repeats

RepeatMasker 36 in conjunction with a curated rice (O. sativa) repeat database was used to identify and classify previously identified repetitive elements. RECON 37 was used to identify repeats de novo. Results reported here relate to the number of bases covered by each element class, rather than the total number of members of each element class with the exception of small transposable elements (that is, miniature inverted repeat transposable elements (MITEs) and short interspersed elements (SINEs)), as BES reads are too short to span complete elements. The repeat content results ranged from a high of 76% in O. australiensis [EE] to a low of 36% in O. coarctata [HHKK] (Table 4). The repeats extracted by RECON overlapped with those identified by RepeatMasker by 32% in O. coarctata [HHKK] to 74% in O. punctata [BB]. This is expected as the curated repeat library used for RepeatMasker is O. sativa [AA] specific and, therefore, more distantly related species would be expected to share fewer repeats. Interestingly, the total repeat content of two CC genome containing allotetraploid species (O. minuta [BBCC] and O. alta [CCDD]) was less than the CC diploid genome of O. officinalis. It was also observed that the highest amount of unique or species-specific repetitive DNA was present in O. australiensis [EE] (21% corresponding to 203 Mb) and the lowest in O. granulata [GG] (10% corresponding to 92 Mb). These sequences were compiled into species-specific repeat databases.

Table 4

Repeat content analysis of 12 Oryza species using RepeatMasker and RECON

Repeat content by RECON

Species

Repeat content by RepeatMasker

Total

Overlap with RepeatMasker (% of total)

Uniqe (% of total)

Total repeat content

O. sativa (Nipponbare)

29.4%

37.2%

23.8% (64%)

13.4% (36%)

42.8%

O. nivara

37.0%

47.6%

29.5% (62%)

18.1% (38%)

55.1%

O. rufipogon

36.6%

49.7%

34.6% (70%)

15.2% (30%)

51.7%

O. glaberrima

28.5%

30.9%

19.8% (64%)

11.2% (36%)

39.7%

O. punctata

40.8%

41.4%

30.7% (74%)

10.7% (26%)

51.5%

O. officinalis

47.2%

56.5%

37.6% (67%)

18.9% (33%)

66.1%

O. minuta

44.0%

46.3%

32.2% (70%)

14.1% (30%)

58.1%

O. alta

43.2%

47.9%

33.2% (69%)

14.7% (31%)

57.9%

O. australiensis

55.0%

66.4%

45.4% (68%)

21.0% (32%)

76.0%

O. brachyantha

20.9%

27.5%

11.6% (42%)

15.9% (58%)

36.8%

O. granulata

30.1%

30.9%

20.5% (66%)

10.4% (34%)

40.5%

O. ridleyi

35.6%

42.7%

23.0% (54%)

19.7% (46%)

55.2%

O. coarctata

19.7%

24.5%

7.7% (32%)

16.7% (68%)

36.4%

The amount of repeats in each genome was directly proportional to the genome size except for O. granulata [GG] (Additional data file 3), suggesting that repetitive sequences play a major role in genome size expansion in the genus Oryza. O. australiensis [EE] had a high amount of repetitive DNA, consistent with the observation of a rapid burst of retrotransposons in this species 38.

We attempted to sub-classify the repetitive elements identified by their class. Only the results from the RepeatMasker search were used for classification (Figure 2) as the de novo repeats identified by RECON were mostly truncated. Retrotransposons (both long terminal repeats (LTRs) and non-LTRs) comprised the major fraction of the total repeats in all species and ranged from 46% of the genome in O. australiensis [EE] to 9% in O. brachyantha [FF]. Among all types of retroelements, the Ty-3/gypsy elements were the most predominant, followed by Ty-1/copia, LINEs and SINEs, respectively. The second most abundant class was DNA (non-MITE) transposons consisting mainly of the cacta-En/Spm, mutator, mariner, ping/pong/snoopy, hAT, and Ac/Ds elements, with En/Spm being the most predominant. These non-MITE DNA transposons ranged from 7% of the genome in O. ridleyi [HHJJ] to 2% in O. coarctata [HHKK]. An exception to this trend was O. brachyantha [FF], which had more MITEs (7%) than DNA (non-MITE) transposons (2%). MITEs were the third most abundant elements in the genus Oryza with the exception of O. brachyantha; MITEs ranged from 0.9% of the genome in O. australiensis to 7% in O. brachyantha. The three AA genome species, as well as O. sativa, have approximately equal amounts of both DNA (non-MITE) transposons as well as MITEs, although with slight differences. Since MITEs represent a large class of repeats with many different types, we calculated the percentage of different MITEs from the overall total as shown in Additional data file 4. Some types of MITEs were more abundant than others and also some were specific to a particular species - for instance, the three AA genome species illustrated a similar pattern for each element. Similarly, the distribution pattern of the MITE elements to the classification of MITE types and their proportion was more comparable among O. punctata [BB], O. officinalis [CC], O. minuta [BBCC] and O. alta [CCDD] genome species rather than those of the other genomes. A similar pattern was also detected between O. ridleyi [HHJJ] and O. coarctata [HHKK] species. This result indicates that the abundance and composition of the MITE elements are closely related to genome type evolution in the genus Oryza.

Figure 2

Comparison of classified repeat composition in the genomes of 12 Oryza species and the O. sativa RefSeq

Comparison of classified repeat composition in the genomes of 12 Oryza species and the O. sativa RefSeq.

Analysis of O. alta [CCDD] revealed that approximately 9% of the BESs were categorized as unclassified repeats, high relative to other 12 Oryza species (2% in O. minuta [BBCC] to 0.2% in O. coarctata [HHKK]), implying the existence of one or more unique classes of repeats in O. alta or the presence of highly diverged repeats that were recognized by our current stringency parameters. On closer inspection, we identified a copia type LTR retrotransposon in a fully sequenced BAC (OA_BBa0237I11) representing a large family of retrotransposons that accounts for the majority of these unclassified repeats. This element, considered in detail elsewhere 39, accounted for two-thirds (2,939,525 bases of 4,392,322 bases) of these unclassified repeat sequences when assayed using cross_match alone.

Simple sequence repeats

SSRs are extremely useful for detecting variation between species and generating molecular genetic maps. A total of 16,980 non-redundant SSRs were detected from the 1.45 million BESs analyzed, which equates to approximately 18 SSRs/Mb of Oryza sequence. Overall, the AA and FF genome species (O. nivara [AA], O. rufipogon [AA], O. glaberrima [AA] and O. brachyantha [FF]) showed the highest SSR content, with an average of 27 SSRs/Mb, whereas the tetraploid (O. minuta [BBCC], O. alta [CCDD], O. ridleyi [HHJJ] and O. coarctata [HHKK]) and non-AA genome diploid species (O. punctata [BB], O. officinalis [CC], O. australiensis [EE], and O. granulata [GG]) showed 17.6 and 17.2 SSRs/Mb, respectively (Table 5). Among the 16,980 non-redundant SSRs, di- and tri-nucleotide SSRs were the most abundant, representing 48.6% and 26.7% of the total dataset, respectively. The relative frequency of di-nucleotide SSRs in all the Oryza species was 6-7% higher than in the O. sativa reference genomes (Table 5; Additional data file 5). In contrast, the relative frequency of tri-nucleotide SSRs was 4-5% lower than in the O. sativa reference genomes and was inversely related to that of the di-nucleotide repeats in each species (Additional data file 5).

Table 5

Distribution of OMAP non-redundant SSRs by motif type

Non-redundant

Tri

Tetra

Penta

Hexa

Total

Species

Ploidy

Total no. of SSRs containing BESs (%)

No.

SSR density (no. of SSRs/Mb)

O. sativa spp japonica

8,636

41.81

6,259

30.31

2,439

11.81

2,628

12.72

691

3.35

20,653

55.8

O. sativa spp indica

8,317

42.94

6,059

31.28

2,285

11.80

2,142

11.06

567

2.93

19,370

47.2

O. nivara

2,129 (2.01)

739

46.33

451

28.28

189

11.85

170

10.66

2.88

1,595

22.5

O. rufipogon

1,599 (2.25)

696

49.40

375

26.61

169

11.99

137

9.72

2.27

1,409

28.2

O. glaberrima

1,289 (1.93)

518

45.00

319

27.72

124

10.77

153

13.29

3.21

1,151

29.5

O. punctata

1,043 (1.53)

408

44.54

268

29.26

105

11.46

100

10.92

3.82

916

18.7

O. officinalis

1,397 (1.35)

545

47.43

308

26.81

111

9.66

146

12.71

3.39

1,149

16.0

O. australiensis

1,306 (0.96)

597

54.17

230

20.87

8.53

128

11.62

4.81

1,102

13.0

O. brachyantha

1,630 (2.42)

590

48.68

287

23.68

135

11.14

166

13.70

2.81

1,212

26.9

O. granulata

1,174 (0.85)

603

55.99

313

29.06

5.94

6.78

2.23

1,077

11.6

O. minuta

1,939 (1.14)

921

54.27

386

22.75

176

10.37

168

9.90

2.71

1,697

17.9

O. alta

1,582 (1.23)

812

52.83

363

23.62

182

11.84

141

9.17

2.54

1,537

20.5

O. ridleyi

2,305 (1.13)

709

35.36

741

36.96

238

11.87

257

12.82

2.99

2,005

15.5

O. coarctata

2,671 (1.37)

1,108

52.02

500

23.47

257

12.07

198

9.30

3.15

2,130

16.5

OMAP total