According to an article, in 2005, by Nicholas Wade of The New York Times, the notion that human evolution had effectively stalled in the distant past (i.e., 50,000 years ago) had been widely accepted.  As recently as 2007, the eminent Harvard psychologist, author, and advocate for evolution, Steven Pinker, revealed, in the context of announcing a change in his views, that he had believed for many years that human evolution had ceased some thousands of years earlier by the time of the agricultural revolution.  Although I have previously argued (2008) that neither first principles nor the totality of available evidence (even some years back) offer much support for this concept of human evolution having ground to a halt, there have been significant challenges in devising reliable methods by which DNA sequence information alone could be used to infer the action of selection on particular genes or regions of the genome.

A new study from Grossman et al., published online in Sciencexpress, reports a novel approach to identifying causal genetic variants responsible for positive selection in human (and possibly other) genomes.  Actually, Sabeti and colleagues describe a fusion of three pre-existing methods as well as two new techniques for identifying such genetic variants.   The five calculated indices are based on three distinct patterns of genetic variation that tend to be associated with selection.  First, alleles that increase in frequency rapidly can have longer-than-average haplotype lengths, meaning that non-random associations between the selected allele and alleles at other loci in the chromosomal vicinity of the locus where the favored allele resides will tend to extend over longer distances than would be the case in the absence of selection.  Second, alleles newly created by mutations in pre-existing alleles (so-called derived alleles) and that confer reproductive advantages will tend to increase in frequency more than can be explained by genetic drift.  This process will also tend to increase the frequencies of other closely linked derived alleles (i.e., derived alleles at nearby loci).  Third, the authors make use of the possibility that derived alleles will be at higher frequency in some geographically-limited populations than in others due to geographically-limited selection, so-called differentiated alleles.

After describing the details of the five measures of selection that they combined into one statistic termed the Composite of Multiple Signals (or CMS), Grossman et al. performed a series of simulations under multiple plausible scenarios for human population structure and verified that CMS was superior to any single statistic in achieving both localization of the selection signal and identifying the causal variant.  They then applied their composite measure to empirical data relating to 185 regions of the human genome identified as subject to recent positive selection by previous analyses associated with the HapMapII project.  Using CMS, the authors were able to identify both known and previously unrecognized variants that were likely to be under recent positive selection.  For genomic regions previously suspected of containing causal variants, the CMS statistic improved spatial localization and causal variant identification.

It will be of interest to follow future applications of CMS as well as to find out what percentage of selection signals identified by CMS can be verified by independent experimental means.  The authors acknowledge that even with this apparently improved approach to analyzing DNA sequence data, fuller understanding of instances of human selection will often require correlated functional data on alternative candidate variants.

References

Wade, N. Brain May Still Be Evolving, Studies Hint. The New York Times, September 9, 2005. http://www.nytimes.com/2005/09/09/science/09brain.html?scp=165&sq=nicholas+wade&st=nyt

Pinker, S. Have Humans Stopped Evolving? http://www.edge.org/q2008/q08_8.html

Greenspan, N.S. Darwin and Deduction. The Scientist. May 9, 2008.http://www.the-scientist.com/templates/trackable/display/news.jsp?type=news&o_url=news/display/54632&id=54632

Grossman SR, Shylakhter I, Karlsson EK, Byrne EH, Morales S, Frieden G, Hostetter E, Angelino E, Garber M, Zuk O, Lander ES, Schaffner SF, Sabeti PC. A Composite of Multiple Signals Distinguishes Causal Variants in Regions of Positive Selection. Science 2010 Jan 7. [Epub ahead of print]http://www.sciencemag.org/cgi/rapidpdf/science.1183863v1.pdf