Primate neurogenomics

James Sikela has a review in PLoS Genetics on genomics as applied to understanding human cognitive evolution (via Brainethics).

I think the article is really interesting, both as a review (if you're interested in the field), and as an illustration of the real limitations of genomics to figure any of this stuff out.

The abstract:

The recent publication of the initial sequence and analysis of the chimp genome allows us, for the first time, to compare our genome with that of our closest living evolutionary relative. With more primate genome sequences being pursued, and with other genome-wide, cross-species comparative techniques emerging, we are entering an era in which we will be able to carry out genomic comparisons of unprecedented scope and detail. These studies should yield a bounty of new insights about the genes and genomic features that are unique to our species as well as those that are unique to other primate lineages, and may begin to causally link some of these to lineage-specific phenotypic characteristics. The most intriguing potential of these new approaches will be in the area of evolutionary neurogenomics and in the possibility that the key human lineage-specific (HLS) genomic changes that underlie the evolution of the human brain will be identified. Such new knowledge should provide fresh insights into neuronal development and higher cognitive function and dysfunction, and may possibly uncover biological mechanisms for information storage, analysis, and retrieval never previously seen.

The first half of the paper reviews several aspects of human genomic structure compared to that of chimpanzees, and to a lesser degree macaques.

The emphasis (although not in so many words) is on illustrating the ways that a genome-centered approach differs from a gene-centered approach. We do not necessarily have access to genetic changes (yet) but we do have access to broad structural changes in the human and chimpanzee genomes, hints about the frequency of certain kinds of changes (duplications and insertions, for example) and what classes of genetic changes have been most frequent.

But there is a real disconnect between this information from genomics and the information that might be relevant to uncovering the evolutionary histories and functions of primates. This is sobering:

Finally, just as having genome sequences available from several different primates will make it possible to more confidently identify HLS genomic changes, the same will be true for each of the individual primate lineages for which genome-wide data will be available. As a result, we can expect to see numerous new discoveries that identify genomic changes specific to each of these primate species. It would therefore seem to be an opportune time to establish programs aimed at sorting out how such changes relate to phenotypic differences among these lineages.

The thing is, we have only the faintest clue how to proceed toward this understanding for humans, whose phenotype we know in excruciating detail. For most other primates, the "phenome" (as the full gamut of the phenotype has lately been called) approximates a black hole. We know a lot about other primates, of course -- a more or less complete survey of their diets, anatomies, and social behaviors for a start. But when did these features come about? What ecologies do they address? Why do primate species transit from one character state to another? And what is left that we don't know? We know orders of magnitude more about humans than about any other primate.

How might we go about matching the specific genomic changes in each primate species with their phenotypes? Maybe we will be very lucky, and we will find a few genes with recurring patterns of evolution in different lineages that share phenotypes -- for instance, there might be a few "sexual dimorphism" genes, or "male dispersal" genes.

But there will be a large thicket of genetic changes in each species. Which genetic changes are fundamental, and which are fine-tuning that happens after more basic changes? Do different species tend to find different genetic mechanisms for parallel features or common ones? Were different primate lineages constrained by the genomic options of their ancestors, or did they evolve readily from one form to another? How many ways are there to achieve functional transitions, such as changes in dispersal patterns or group size?

To some extent, we can bracket these probabilities. There are more transitions in favored group size among primates than transitions in locomotor adaptations, for instance. Likewise, when we look more broadly at other mammals, there are many more transitions in group size than in locomotor behavior. So we might guess that locomotor anatomy will be more constrained than sociality. We might also guess that transitions in locomotor anatomy will be more likely to involve a few critical genes again and again in different mammal and primate lineages, while changes in sociality may have involved large numbers of genes, and different sets in different evolving lineages.

Of course, the aim of the current paper is to outline ways of finding genetic changes underlying human cognitive evolution. And since some cognitive features of humans are not shared by any other animal, and other cognitive features that are shared with other primates are shared quite broadly (and discontinuously) with other primates and other mammals, the genetic story probably involves many genes that will not show clear signatures of evolutionary change in other lineages -- as well as many others that do show changes in one or more primates or mammals.

The approaches described in the paper for finding important genes underlying human cognitive evolution really illustrate the limits of this science right now. There are essentially three strategies:

  1. Focus on genes that are brain-expressed and vary a whole lot between humans and chimpanzees (or other primates).
  2. Focus on genes that are involved in genetic disorders of cognition.
  3. Try to make transgenic animal models.

Now, really, these are the approaches used in most evolutionary genomics and evolutionary developmental biology right now. But cognition generates some problems that aren't so pressing for most other topics. For developmental genes involved in the anatomy of locomotion, for example, there are clear animal homologues for the human pelvis, femur, tibia, foot, and so on. Fields of action of genes can be compared in different species, and -- even though little experimentation in humans is possible -- reasonable hypotheses for the effects of genetic changes in humans can be inferred. And knockout mice missing a gene will show obvious effects if that gene is necessary for the early development of a femur, as will humans with congenitally missing anatomical features.

But many human cognitive features are abstracted far from anatomy. What does it mean to take FoxP2 away from a mouse? Humans with a mutant version of the gene have specific language impairment, but what anatomical features are different? How do the neural pathways necessary to language even develop?

The most complicated problem is not that humans will be so very different from other primates. Certainly humans are very different, but if we can arrive at hypotheses for the origin of human cognitive characters as a system, we will find ways to test them -- whether with genes, bones, or stone tools.

To my mind, the most complicated problem is that humans may not be very different at all in their evolutionary pathways from other primates. If the changes underlying human cognitive evolution involved a similar number or scope of genetic changes as those underlying the cognitive evolution of other primates -- like capuchins, for instance --- it may be extremely hard to figure out how broadly similar evolutionary pathways came to such different outcomes. And if the genetic systems are at all complex, there may be little hope of figuring out how any non-human primate (with no behavioral record from the past) came to its current species-specific cognitive niche.

Maybe we are left hoping for an unexpectedly simple system. There was little hope that anyone would figure out evolution of body form in different phyla of animals -- and then they discovered the Hox genes, a simple, almost mechanistic, system controllling segmentation and many downstream components of development. Will there be such a system for cognitive outputs of the mammalian brain?

In any event, the hard part is working from the phenotype -- behavior and anatomy -- back to the genome. And working from the ecology, which ultimately underlies reproductive success and evolutionary history, to the phenotype. Not only which anatomies and behaviors are important, but when, and how are they genetically or environmentally influenced. And how do we parse these facts among the thousands of genetic changes that might explain them?


Sikela JM (2006) The Jewels of Our Genome: The Search for the Genomic Changes Underlying the Evolutionarily Unique Capacities of the Human Brain. PLoS Genet 2(5): e80 Online free at PLoS