Positive Natural Selection in theHuman LineageP. C. Sabeti,1,2*S. F. Schaffner,1*† B. Fry,1J. Lohmueller,1,3P. Varilly,1O. Shamovsky,1A. Palma,1T. S. Mikkelsen,1D. Altshuler,1,4,5E. S. Lander1,6,7,8Positive natural selection is the force that drives the increase in prevalence of advantageous traits,and it has played a central role in our development as a species. Until recently, the study of naturalselection in humans has largely been restricted to comparing individual candidate genes totheoretical expectations. The advent of genome-wide sequence and polymorphism data bringsfundamental new tools to the study of natural selection. It is now possible to identify newcandidates for selection and to reevaluate previous claims by comparison with empiricaldistributions of DNA sequence variation across the human genome and among populations. Theflood of data and analytical methods, however, raises many new challenges. Here, we reviewapproaches to detect positive natural selection, describe results from recent analyses of genome-wide data, and discuss the prospects and challenges ahead as we expand our understanding of therole of natural selection in shaping the human genome.Homo sapiens, like all species, hasbeen shaped by positive natural selec-tion. As first articulated by Darwinand Wallace in 1858, positive selection is theprinciple that beneficial traits—those that makeit more likely that their carriers will survive andreproduce—tend to become more frequent inpopulations over time (1). In the case of hu-mans, these beneficial traits likely includedbipedalism, speech, resistance to infectious dis-eases, and other adaptations to new and diverseenvironments. Understanding the traits (andgenes underlying them) that have undergonepositive selection during human evolution canprovide insight into the events that have shapedour species, as well as into the diseases thatcontinue to plague us today.Until very recently, the only practical wayto identify cases of positive selection in hu-mans was to examine individual candidate genes.Allison noted in 1954 that the geographicaldistribution of sickle cell disease was limitedto Afric a and correlated with malaria endemic-ity (2); this observation led to the identificationof the sickle cell mutation in the Hemoglobin-Bgene (HBB) as having been the target of selec-tion for malaria resistance (3, 4). Since then,approximately 90 different loci have been pro-posed as possible targets for selection (table S1provides a review of this literature).Some of the proposed candidates for se-lection, like HBB, have strong support in theform of a functional mutation with an identifiedphenotypic effect that is a likely target of se-lection. In the case of HBB, the selected muta-tion creates a glutamatetovalineaminoacidchange, but the target of selection need not bein the protein-coding region of a gene. Forexample, the Duffy antigen (FY) gene encodesa membrane protein used by the Plasmodiumvivax malaria parasite to enter red blood cells.A mutation in the promoter of FY that disruptsprotein expression confers protection againstP. vivax malaria and was proposed to be se-lected for in regions of Africa where P. vivaxmalaria has been endemic (5). Another exampleis a mutation in a regulatory region near thegene for lactase (LCT ) that allows lactose tol-erance to persist into adulthood. This particularvariant was apparently selected in parts ofEurope after the domestication of cattle (6).Often, however, the functional target of se-lection is not known. In some cases, candidategenes gain support because they lie in func-tional pathways, such as spermatogenesis andthe immune response, that are known to be fre-quent targets for selection in other species. Oneexample is protamine 1 (PRM1), a sperm-specificprotein that compacts sperm DNA (7, 8). Suchcases, however, are the exception. Most pro-posed candidates lack compelling biologicalsupport. Rather, the argument for selection hasrelied solely on comparative and populationgenetic evidence.Despite its great potential to illuminate newbiological mechanisms, identification of se-lected loci by genetic evidence alone is fraughtwith methodological challenges. Studies basedon comparisons between species suffer fromlimited power to detect individual incidents ofselection, whereas studies based on humangenetic variation have suffered from difficultieswith assessing statistical significance. The evi-dence for positive selection has traditionallybeen evaluated by comparison with expecta-tions under standard population genetic models,but the model parameters (especially those re-lating to population history) have been poorlyconstrained by available data, leading to largeuncertainties in model predictions. One solutionwould be to assess significance by comparingempirical results from different studies, butthis has been challenging because of the variedstatistical tests, sizes of genomic region, andpopulation samples used (see table S2 forexamples).The advent of whole-genome sequencingand increasingly complete surveys of geneticvariation represent a turning point in the studyof positive selection in humans. With these ad-vances, humans can now join model organismssuch as Drosophila (9) at the forefront of evo-lutionary studies. Newly available tools allowsystematic survey of the genome to find thestrongest candidate loci for natural selection, aswell as to reevaluate previously proposed can-didate genes, in comparison with genetic varia-tion in the genome as a whole (the genome-wideempirical distribution). Although they permit usto make progress even while working out re-maining theoretical issues, they also bring ana-lytical challenges of their own, because theyrepresent imperfect samples of genetic variation.Here, we review genetic methods for de-tecting natural selection, discuss initial resultsabout positive selection based on recent whole-genome analyses, and outline the potential andthe challenges ahead in going from candi-dates of selection to proven examples of adapt-ive evolution.Methods for Detecting SelectionWhen alleles (genetic variations) under positiveselection increase in prevalence in a population,they leave distinctive ‘‘signatures,’’ or patternsof genetic variation, in DNA sequence. Thesesignatures can be identified by comparison withthe background distribution of genetic variationin humans, which is generally argued to evolvelargely under neutrality (10). This is in accordwith the neutral theory,
View Full Document
Unlocking...