Now that the sequence of the human genome is almost complete, the human genomics community is turning its attention towards what, genetically speaking, makes people different.
In one of the most considerable studies of human genetic variation so far, scientists at Perlegen Sciences and their collaborators have quantified the variation at more than 1.5 million DNA positions in 71 individuals of European American, African American and Chinese ancestry. Their findings are described in Science1. Not too long ago, genetic inferences about human ancestry were based primarily on a single gene2; now we have data on variants from millions of positions in the human genome that may help us to uncover human genetic ancestry and realize the promise of ‘personalized’ medicine.
There are roughly 10 million positions in the human genome that vary among individuals. To a great extent, the genetic causes for differences in physical and mental characteristics — including susceptibility to certain diseases — will probably be found in these ‘single nucleotide polymorphisms’ (SNPs; Fig. 1). The focus of large-scale genomics studies is now being turned towards identifying and describing this variation, one goal being to develop new treatment options for many diseases. In the process, we are learning much about our evolutionary and demographic history, through statistical analyses of the current genetic make-up of human populations.
Hinds et al.1 provide the first comprehensive study of human DNA variation in which the same SNP ascertainment scheme was used throughout the genome. An ascertainment scheme is the method by which SNPs are chosen. It effectively provides a filter that influences the patterns observed in the data3 — much like the colour of the shades in sunglasses selectively filters out certain wavelengths of light, changing the colour of the sky for the person using them. In many large-scale studies of human genomic variation, including the public HapMap project4, different ascertainment schemes have been used in different genomic regions, complicating comparisons among regions. But Hinds et al. chose a scheme that is uniform among genomic regions for the vast majority of SNPs. Their data are therefore perfectly suited for identifying genomic regions with reduced levels of variability, regions that are enriched in population-specific differences, and regions that have undergone natural selection.
Hinds et al. find that correlation among different genetic variants — or ‘linkage disequilibrium’ — is enhanced in regions that contain genes as opposed to non-coding regions of DNA, and among SNPs that cause amino-acid changes in an encoded protein. This observation provides indirect evidence for the action of natural selection on at least some of the variants in the coding regions. Many factors have influenced human genetic variation, including demographic processes of population growth, subdivision and migration. However, some genes have also been targeted by natural selection as humans have adapted to new environments, fought pathogens5, and developed advanced cognitive abilities and culture6. If selection has occurred very recently, it leaves a distinct pattern in the genome — for example by increasing the frequency of particular variants or the degree of genetic differentiation among populations — and this pattern can be distinguished from the background pattern of variation using statistical methods. Further statistical analyses of Hinds and colleagues' data may identify the specific genes or regions that have been targeted by natural selection in the recent evolutionary history of mankind.
Such studies are of interest not only because they may help to elucidate the molecular forces underlying human evolution, but also because regions targeted by selection may be associated with genetic diseases or disease susceptibility7. Disease-causing mutations are likely to be selected against, as long as the disease reduces the fertility or viability of the affected individual. However, some genetic-disease factors may also have increased in frequency in the population because they give the carriers a fitness advantage8,9,10. A classic example is mutations in the β-globin gene, in which individuals carrying two copies of a disease mutation develop sickle-cell anaemia, but individuals with one copy8 have partial resistance against malaria. It has also been proposed that the mutations that cause cystic fibrosis9 and Tay–Sachs10 disease might reduce susceptibility to typhoid and tuberculosis, respectively, in carriers with only one copy of the mutation. Detailed statistical and evolutionary analyses of Hinds and colleagues' data may suggest which other genomic regions, or genes, are most likely to harbour disease-associated mutations, even without the help of any data on the consequences of the mutations.
The availability of large-scale SNP data therefore brings together the disciplines of medical and evolutionary genetics. Evolutionary thinking underlies many of the common methods used for identifying associations between genetic types and observable traits or diseases using population genetic data, and has led to major advances in genetic epidemiology11,12. And with the increased emphasis on identifying the determinants of genetic diseases comes the awareness that human genomic variation can only be understood fully in light of the evolutionary forces that have shaped it. The data published by Hinds et al.1 will provide a unique resource for explorations in both disciplines.
Hinds, D. A. et al. Science 307, 1072–1079 (2005).
Cann, R. A., Stoneking, M. & Wilson, A. C. Nature 325, 31–36 (1987).
Nielsen, R., Todd, M. J. & Clark, A. G. Genetics 168, 2373–2382 (2004).
Gibbs, R. A. et al. Nature 426, 789–796 (2003).
Sawyer, S. L. et al. PLoS Biol. 2, e275 (2004).
Enard, W. et al. Nature 418, 869–872 (2002).
Halushka, M. K. et al. Nature Genet. 22, 239–247 (1999).
Friedman, M. J. & Trager, W. Sci. Am. 244, 154–164 (1981).
Pier, G. B. et al. Nature 393, 79–82 (1998).
Rotter, J. I. & Diamond, J. M. Nature 329, 289–290 (1987).
Slatkin, M. Genetics 137, 331–336 (1994).
Pritchard, J. K., Stephens, M., Rosenberg, N. A. & Donnelly, P. Am. J. Hum. Genet. 67, 170–181 (2000).
About this article
Annual Review of Genomics and Human Genetics (2006)