Ancestral primate viewed

Article metrics

How did the genomes of modern mammals come to be organized the way they are? Improvements in genome sequencing, gene mapping and chromosome recognition mean that, by comparing the order and sequence of genes in different genomes, we may be able to trace the evolutionary pathways that our ancestors took. The potential of such an approach is illustrated by three new studies1,2,3, which compare the genome maps and chromosomes of 15 species of primate, using four non-primate orders — carnivores, rodents, artiodactyls (hoofed mammals) and tree shrews — as a reference standard for the organization of ancestral mammalian genomes. Although the approach itself is not new, the chance to explore entire vertebrate genomes with such a high level of precision is.

Comparative cytogenetics (the study of differences in chromosome structure and appearance) dates back three decades. Then, factors such as the position of the centromere, a chromosomal structure involved in nuclear division, alerted researchers to differences between distantly related species4. Later, specific banding patterns (obtained by staining chromosomes) were used to infer sub-chromosomal homologies, and many of these have turned out to be correct5. Today, comparative gene maps have been started for over 40 mammalian species6. Chromosome painting (where flow-sorted individual chromosomes of one species are labelled with fluorescent dyes and hybridized in situ to chromosomes of another species) has allowed long stretches of conserved gene segments to be identified among very distantly related species.

Such comparisons have given us a glimpse of the genome rearrangements that punctuate adaptation and species formation7. Because exchanges between genomes are rare, and the junctions between gene segments can be easily identified, such rearrangements are powerful evolutionary characters that proscribe the history of species that retain them. Comparative ‘phylogenomic’ approaches therefore combine what we know about segment homology with the strategies of evolutionary cladistics (a method of classification based on those shared characteristics that are assumed to indicate a common ancestry). In the first of the new papers1, Haig restates an idea often heard from palaeontologists working with a tooth or cranial fragment, when he writes that describing the process of reconstructing ancestral chromosome structures “can be likened to the difficulty of writing an entertaining account of the piece-by-piece assembly of a jigsaw puzzle”. The process is indeed tedious and treacherous, but the goal of reconstructing genomic history is a lofty one.

So how is it done? First, work out how many chromosomes or segments are conserved between the genomes of two species. Second, from this count, estimate how many chromosome exchanges (translocations, transpositions, fusions, fissions or inversions) it takes to rearrange one genome to the other — that is, how many scissor cuts would allow a human genome arrangement to be turned into that of a tree shrew, and vice versa? Next align additional species to the first two and identify the conserved homology blocks. The ancestral forms can be regarded as those shared between several species. (By contrast, derived segments or connections are unique to one or more closely related species, and serve as evolutionary signatures of particular lineages.) Finally, assemble the hypothetical ancestral genome — for example, for all primates — by comparison with the genomes of ‘outgroup’ species. In the new studies these included tree shrews; carnivores (cats, mink and seals); perissodactyls (horses); artiodactyls (cows, sheep and goats); or rodents (rats and mice). The ancestral chromosomal traits were defined as those shared between primates and the outgroups.

This process allows the minimum-sized fragments of chromosome exchange to be identified (termed the smallest conserved evolutionary unit segment; SCEUS8). The SCEUS is then treated as an evolutionary character. Haig1 introduces a ‘Cambridge grid’, which tabulates the chromosomes of one species along the horizontal axis and those of a second species vertically. He then assigns individual Greek letters in the matrix to ancestral SCEUS associations. Once the ancestral genome is designed, the genomic exchanges that have occurred in lineages leading to the living species can be recapitulated.

The new studies1,2,3 interpret available data to reconstruct chromosomal characteristics from the ancestor of all primates (Fig. 1, overleaf). This creature last lived at least 65 million years ago, so it seems remarkable that, compared to the ancestral types, 18–20 human chromosomes remain unchanged and the rest have but a single exchange. Different exchanges have occurred in the lineages leading to distinct primate families and genera. But, for most species, fewer than 20 rearrangements are enough to re-assort modern genomes to that of the primate ancestor. At least 12 of the ancestral primate chromosomes are also found intact in the genomes of humans, cats, seals, cows and shrews. So the ancestral primate genome must differ only marginally from the ancestral mammalian genome, which pre-dates the origins of carnivores, primates and artiodactyls.

Figure 1: Reconstruction of the ancestral genome of living primates, depicted as conserved and rearranged human chromosomes1,2,3.

To the right of each ancestral primate chromosome is the human chromosome number. To the left are: the number of primate species (out of 15) in which the chromosome synteny is conserved as it occurs in the ancestral primate genome (yellow); the number, out of 15, in which a whole chromosome homology block is retained, but attached to another chromosomal piece (green); and the same counts for 11 outgroup species including pig, cow, muntjac, dolphin, cat, seal, mink, horse and shrew1,2,3 (red and blue). Arrows indicate proposed separations of human chromosomes postulated by Müller et al.3 (chromosome 1) and Haig1 (chromosome 8), which are unique to these authors.

The new papers also affirm that neither the positions of chromosomal breaks nor the evolutionary rate of rearrangements in primates seem to follow predictable patterns. Breaks can occur in many positions on any chromosome. Many species of primate, such as humans, great apes and Old World monkeys, show a remarkably slow rate of genome exchange — on the order of one or two exchanges every ten million years. By contrast, other species (gibbons, owl monkeys and lemurs, for example) globally reorganize their genomes, with two to four times more rearrangement relative to organization of the ancestral genome. A similar rate dichotomy is apparent in other mammalian orders, where a slow/default rate seen in most lineages (such as cat, mink, shrew and pig) is punctuated by drastic genome shuffles in others (dogs, bears, rodents and bears)7.

But there are limitations to what we can infer from comparative gene maps and chromosome painting. First, only a handful of mammalian species (human, mouse, rat, pig, goat, sheep, cow, cat and dog) have comparative maps with enough genes to allow their genomes to be compared6,7. Second, although chromosome painting (on which most of the primate analyses are based) allows whole genomes to be assessed, small segments can be overlooked because of weak DNA hybridization. Precision is improved with reciprocal painting9,10,11, or when more closely related index species are used as probes10,11. Third, the order of the genes within homology segments was not considered in the primate assessment shown in Fig. 1. This means that an important class of chromosome exchanges, the interstitial inversions and translocations, is excluded. Comparative alignments of human gene order with ordered gene maps of non-primate species (goat and cat)12,13 have uncovered two to three times more breaks than were revealed by chromosome painting.

Another hurdle is analytical — there is no consensus algorithm for tracking genomic exchanges. Assumptions about the randomness of genome exchanges are probably over-simplifications, and there may be favoured site changes that introduce confounding convergent changes to evolutionary interpretations. So far, little emphasis has been placed on discordant exchange rates among lineages, and for the different categories of exchange observed. Weighting of such characters in evolutionary analyses is an important consideration that needs theoretical and empirical input. Finally, the choice of species sampled could severely bias interpretations. Discrimination of shared-derived, as opposed to shared-ancestral arrangements, depends on how frequently they occur in the outgroups studied (Fig. 1). More is better in statistical terms, yet only a few species6,7,9 in fewer than half of the mammalian orders have been assessed.

Powerful new genomic and gene-mapping technologies should overcome these limitations. The resolution of comparative genome mapping is approaching its highest power ever, allowing linear maps of index species to be aligned explicitly. The primates offer the first and most accurate look at the history of human genome organization, but certainly not the last. We'll soon have similar reconstruction among other mammalian orders7 and even beyond, as previewed on page 411 of this issue14, where the chicken gene map weighs in on a comparative genomics perspective.


  1. 1

    Haig, D. Phil. Trans. R. Soc. Lond. B 354, 1447–1470 (1999).

  2. 2

    Chowdhary, B. P. et al. Genet. Res. 8, 577–589 (1999).

  3. 3

    Müller, S. et al. Chromosoma 108, 393–400 (1999).

  4. 4

    Hsu, T. C. & Bernirschke, K. (eds) An Atlas of Mammalian Chromosomes Folio 122 1774 (Springer, Berlin, 1969).

  5. 5

    Yunis, J. J. et al. Science 208, 1145–1148 (1980).

  6. 6

    Graves, J. A. M. & Vandeberg, J. Inst. Lab. Anim. Res. J. 39, 47–260 (1999).

  7. 7

    O'Brien, S. J. et al. Science 286, 458–481 (1999).

  8. 8

    O'Brien, S. J. et al. Nature Genet. 3, 103–112 (1993).

  9. 9

    Wienberg, J. & Stanyon, R. Curr. Opin. Genet. Dev. 7, 784–791 (1997).

  10. 10

    Nash, W. et al. Cytogenet. Cell Genet. 83, 182–192 (1998).

  11. 11

    Stanyon, R. Cytogenet. Cell Genet. 84, 150–155 (1999).

  12. 12

    Schibler, L. et al. Genome Res. 8, 901–915 (1998).

  13. 13

    Murphy, W. et al. Genome Res. (submitted).

  14. 14

    Burt, D. W. et al. Nature 402, 411–413 (1999).

Download references

Author information

Correspondence to Stephen J. O'Brien.

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.