Update on Comparative GenomicsBeyond the Arabidopsis Genome: Opportunities forComparative Genomics1Anne E. Hall2, Aretha Fiebig2, and Daphne Preuss*Howard Hughes Medical Institute, The University of Chicago, 1103 East 57th Street, Chicago, Illinois 60637Like most higher eukaryotes, flowering plants arebelieved to contain surprisingly similar numbers ofgenes. Nevertheless, angiosperm genome sizes varyover a wide range—from 50 Mb to over 120,000 Mb.Comparative mapping has shown that numerous al-terations contribute to genomic diversity amongplants. Over time, chromosomes are broken, reas-sembled, partially or wholly duplicated, and eveneliminated, ultimately resulting in reproductive iso-lation and speciation. However, the mechanisms thatcreate such variation, and the evolutionary forcesthat fix these changes, are not well understood. Com-parative analyses of plant genomes promise to clarifythe selective pressures driving these changes; suchinvestigations will elucidate alterations at the level ofwhole genomes, as well as those at the level of spe-cific sequences, including genes, repetitive elements,and other non-coding regions.Although low-resolution genetic maps can identifygross chromosomal alterations, a clear understand-ing of the mechanisms behind these changes requiresmultispecies sequence comparisons. Such analysesreveal the composition, organization, and functionalcomponents of genomes and provide insight intoregional differences in composition between relatedspecies. In addition, sequence comparisons elucidateevolutionary history; for example, the stepwise accu-mulation of nucleotide insertions/deletions (indels)only becomes clear with the analysis of multiple spe-cies. Comparative sequence analysis also aids in geneprediction and sequence annotation, and facilitatesthe identification and definition of regulatory ele-ments, including promoters, enhancers, and tran-scription factor-binding sites (Kent and Zahler, 2000;Koch et al., 2001b).The recent analysis of the sequence of the Arabi-dopsis genome highlighted unexpected aspects of itscomposition, organization, and function (Arabidop-sis Genome Initiative [AGI], 2000). The questionsraised by these observations can best be approachedthrough comparative genomics. For example, al-though Arabidopsis is considered a “true” diploid,its genome has undergone major duplication events,followed by extensive rearrangements and chromo-some fusion and loss, hypothesized to have shiftedthe haploid chromosome number from 4 to 8 andthen to 5 (AGI, 2000; Vision et al., 2000). Interestingly,however, evidence of duplications was not found inthe sequenced portions of the centromere regions. Allfive centromeres contain tracts of unique DNA, in-terspersed with similar types of transposable ele-ments and interrupted by large tandem arrays ofsatellites. Aside from the repetitive sequences, pair-wise comparisons of the centromere regions did notidentify blocks of unique sequence indicative of an-cient duplication events (AGI, 2000).In addition to large segmental duplications on thechromosome arms, Arabidopsis also contains a prev-alence of gene families, many of which are the resultof tandem duplications of individual genes, ratherthan redundancy of entire chromosome segments.Nearly 40% of the predicted genes in the Arabidopsisgenome belong to families that contain more thanfive members (AGI, 2000). Through studies of relatedspecies, it will become possible to discern the timingof genome duplications, the types of DNA elimi-nated, and the mechanisms responsible for rear-rangements and deletions. Moreover, analysis of thegenomes of related species will clarify how genefamilies expand, contract, and diversify. Examinationof relatives containing different types of genome du-plications will also reveal the changes that occur aftersuch events, including mechanisms that are activatedafter large-scale genomic perturbation.Clearly, comparative genomic approaches wouldprovide enormous benefit toward understanding theorigins of the Arabidopsis genome, as well as otherplant genomes. Although previous comparisons withgenes from yeast, flies, worms, and mammals haveprovided functional clues for approximately 70% ofArabidopsis genes, the vast evolutionary distancesthat separate these species restrict such comparisonsto coding regions (AGI, 2000). Even the genomes ofArabidopsis and rice (Oryza sativa), which are sepa-rated by approximately 200 million years (MY), havesubstantially diverged (Wolfe et al., 1989), makingthe incremental stages of evolutionary change diffi-cult to grasp (Goff et al., 2002; Yu et al., 2002).In this Update, we explore the utility of a compar-ative genomics approach that relies on Arabidopsis1This work was supported in part by the Howard HughesMedical Institute and by the National Science Foundation (grantno. MCB 0077854 to A.F.).2These authors contributed equally to the paper.* Corresponding author; e-mail [email protected]; fax 773–702–6648.www.plantphysiol.org/cgi/doi/10.1104/pp.004051.Plant Physiology, August 2002, Vol. 129, pp. 1439–1447, www.plantphysiol.org © 2002 American Society of Plant Biologists 1439and several other species within the Brassicaceaefamily. We discuss discoveries made from prior com-parisons within the family, prospects for additionalgenomic studies, and their potential for significantlyimproving our understanding of plant genomes. Fi-nally, we discuss the community resources requiredto launch an effort with the necessary breadth toaddress a wide range of sequence-based evolutionaryquestions, and suggest a set of candidate species forgenomic comparisons in the Brassicaceae.THE BRASSICACEAE: A USEFUL FAMILY FORCOMPARATIVE GENOMICSArabidopsis is a member of the Brassicaceae fam-ily; the wealth of information and resources providedby the community of Arabidopsis researchers conse-quently provides a well-supported infrastructurethat makes the Brassicaceae ideal for comparativestudies of plant genomes. The Brassicaceae family islarge, encompassing approximately 340 genera andmore than 3350 species, a few of which are shown inFigure 1 (Al-Shehbaz, 1984). These species divergedfrom a common ancestor over a time period of ap-proximately 40 to 50 MY as a result of numerousindependent speciation events (Koch et al., 2001a).Thus, the thousands of extant species provide signif-icant opportunities for investigating the genetic dif-ferences that lead to speciation. In practical terms,several