The utility of single nucleotide polymorphisms in inferences of population history

• DOI: 10.1016/s0169-5347(03)00018-1
• Publication Date: 2003-05-13T04:04:58Z
• Authors: Robb T. Brumfield; Peter Beerli; Deborah A. Nickerson; Scott V. Edwards
• Keywords: SNP

Single nucleotide polymorphisms (SNPs) represent the most widespread type of sequence variation in genomes, yet they have only emerged recently as valuable genetic markers for revealing the evolutionary history of populations. Their occurrence throughout the genome also makes them ideal for analyses of speciation and historical demography, especially in light of recent theory suggesting that many unlinked nuclear loci are needed to estimate population genetic parameters with statistical confidence. In spite of having lower variation compared with microsatellites, SNPs should make the comparison of genomic diversities and histories of different species (the core goal of comparative biogeography) more straightforward than has been possible with microsatellites. The most pervasive, but correctable, complication to SNP analysis is a bias towards analyzing only the most variable loci, an artifact that is usually introduced by the limited number of individuals used to screen initially for polymorphisms. Although the use of SNPs as markers in population studies is still new, innovative methods for SNP identification, automated screening, haplotype inference and statistical analysis might quickly make SNPs the marker of choice. Traditionally, phylogeography has used gene trees of nonrecombining, uniparentally inherited LOCI (see Glossary), such as mitochondrial DNA or the vertebrate Y chromosome, to study the geographical distribution of genetic variation within species [1]. As evolutionary biologists have started to examine variation in recombining, biparentally inherited loci, a natural outgrowth of phylogeography is a shift from gene trees to analyses, based on COALESCENT THEORY, of multi-locus, recombining histories. This new discipline, dubbed historical demography [2,3] or statistical phylogeography [4], is concerned less with gene trees than with estimating population parameters such as genetic diversities, divergence times, growth rates and gene flow between populations. The shift in focus is, in part, a result of recent advances in population genetics, which suggest that, from a statistical standpoint, the ability of single-locus phylogeography to determine the timing of speciation events and the historical demography of populations has been overestimated [3‐7]. The errors surrounding estimates of divergence times, rates of gene flow and population-size changes during speciation are all reduced substantially when information from multiple unlinked loci is combined [8,9]. With the move to analyses of multiple loci, phylogeographers must re-learn an old lesson: that the number of loci required to estimate the preceding parameters with statistical confidence can be soberingly large because of the high stochasticity of the gene tree of any single locus [10]. What is required is a suite of unlinked nuclear genetic markers that can capture a genome-wide picture of the population history [3,11‐14]. The polymerase chain reaction (PCR) as well as fluorescent sequencing and fragment analysis technologies have catalyzed a revolution in the development of genetic markers for the analysis of natural populations. Emphasizing discoveries in nonmodel species, we discuss one emerging marker of great relevance to historical demography: single nucleotide polymorphisms (SNPs).