Detection of Signatures of Positive Selection in Naturally Occurring Genetic Variation

Two different questions can be addressed by using molecular markers for making inferences on the effects of positive natural selection. First, when positive selection is proposed but direct evidence is lacking, genetic inference can reveal its existence. Alternatively, when positive selection is evident (fitness advantage is apparent), genetic or genomic tools can be used to identify specific genes or genomic regions that underwent selection. In this review I discuss the methods and progress in detecting the signatures of selection by single-locus, multi-locus and genome-wide analyses. I emphasize 1) the differences between the two approaches, and 2) importance of verification of putative evidences of selection obtained with molecular markers by direct tests of fitness. Testing for adaptation with molecular markers engages primarily two types of positive natural selection that both increase frequency of a favorable allele, either to fixation (directional selection) or to an intermedium equilibrium (balancing selection). Therefore I discuss how signatures of balancing vs. directional selection can be detected by different statistical tests