Evidence that rice and other cereals are ancient aneuploids

Abstract

Detailed analyses of the genomes of several model organisms revealed that large-scale gene or even entire-genome duplications have played prominent roles in the evolutionary history of many eukaryotes. Recently, strong evidence has been presented that the genomic structure of the dicotyledonous model plant species Arabidopsis is the result of multiple rounds of entire-genome duplications. Here, we analyze the genome of the monocotyledonous model plant species rice, for which a draft of the genomic sequence was published recently. We show that a substantial fraction of all rice genes ( approximately 15%) are found in duplicated segments. Dating of these block duplications, their nonuniform distribution over the different rice chromosomes, and comparison with the duplication history of Arabidopsis suggest that rice is not an ancient polyploid, as suggested previously, but an ancient aneuploid that has experienced the duplication of one-or a large part of one-chromosome in its evolutionary past, approximately 70 million years ago. This date predates the divergence of most of the cereals, and relative dating by phylogenetic analysis shows that this duplication event is shared by most if not all of them.

Figure 1.

Overview of the Genomic Scaffolds Generated by ASGAR. (A) Scatterplot showing the number of genes versus the scaffold length for all 966 genomic scaffolds that were used for the detection of duplicated blocks. The best-fit line, which shows a quite homogeneous gene density for the scaffolds (R² = 0.85), represents a gene density of 1 gene per 10 kb. (B) Length distribution of all genomic scaffolds that were subjected to block detection. The line indicates the relative (cumulative) contribution of the scaffolds assigned per bin (i.e., length segment) in the histogram.

Figure 2.

Scheme of Nonhidden, Hidden, and Ghost Duplications. Boxes represent the genes on chromosomal segments of genomes A and B, whereas connecting lines indicate the anchor points (i.e., homologous or duplicated genes). Hidden duplications are heavily degenerated block duplications that cannot be observed by directly comparing the duplicated segments; rather, they are observed only through comparison with a third segment from the same genome. Because nonhidden duplications are used to infer hidden duplications, no additional genomic segments are assigned to a duplication event, although the number of duplication events for a given segment increases. Ghost duplications are hidden block duplications that can be identified only through colinearity with the same segment in a different genome. In contrast to hidden duplications, the identification of ghost duplications increases the fraction of the genome involved in a duplication event.

Figure 3.

Absolute Dating of Block Duplication Events in the Rice Genome. Age distribution of all gene duplicates that are part of large (more than five anchor points) duplicated segments in the rice genome. The line indicates the relative (cumulative) contribution of the anchor points assigned per bin (i.e., age segment) in the histogram.

Figure 4.

Dating of Duplication Events in the Rice Genome by Phylogenetic Means. Expected tree topology and date of origin for genes of the cereals wheat, barley, rice, maize, and sorghum if these genes have duplicated before the divergence of rice and other cereals. The large majority of tree topologies obtained in this study, including those of two copies of rice (i.e., the retained duplicates found in large duplicated segments) and at least one copy of another cereal, are congruent with this tree topology, in which one rice gene branches off before the divergence of rice and other cereals. Such topologies suggest a duplication before the divergence of rice, barley, wheat, maize, and sorghum, estimated at ∼50 million years ago (Kellogg, 2001), and may have occurred just before the origin of the grasses, as suggested by the K_s-based dating (see text for more details).

Figure 5.

Frequency Distribution of Duplicated Genes in Arabidopsis and Rice as a Function of the Number of Silent Substitutions per Silent Site. All frequencies were corrected for the total number of dated gene duplicates per genome, which were 4928 for Arabidopsis (white squares) and 7698 for rice (gray diamonds). The fact that the total number of duplicated genes is higher in the rice than in the Arabidopsis gene family is attributable to the facts that the rice genome contains more predicted genes and that in Arabidopsis more gene families with >10 members have been omitted from the analysis.

Figure 6.

Set of Homologous Chromosomal Segments (Multiplicon) of Arabidopsis and Rice. Arrows represent the genes on the chromosomal segments, and connecting lines indicate the anchor points (i.e., homologous or duplicated genes) that are part of a significant colinear relation determined by the ADHoRe algorithm. For each genomic segment, the names of the two genes delineating the segment are shown. Chromosomal segments of rice and Arabidopsis are shown in gray and white, respectively. By considering the colinearity between Arabidopsis and rice, a set of seemingly unrelated Arabidopsis segments can be joined into a multiplicon with a multiplication level of five, confirming the three duplication events in Arabidopsis described previously (Simillion et al., 2002). This colinearity also reveals that all three rice segments are linked with each other by two duplication events. Scaffold Os04_R2_9 includes BACs with accession numbers AL663006, AL662998, AL606459, AL607006, AL606728, AL606695, AL606587, AL606647, AL606633, AL663000, AL731613, AL606682, AL606687, AL606694, AL606628, AL607001, AL663003, and AL662954; scaffold Os10_5 includes BACs with accession numbers AC084763, AC079890, AC079874, AC069300, AC037426, and AC026758; and scaffold Os02_R2_44 includes BACs with accession numbers AP005108, AP004037, AP004883, AP005072, AP005289, AP005006, and AP004676.