Genetic linkage and molecular evolution

The rate of genetic recombination varies among species, and across different regions of the genome within a species. In the extreme case, all genes in an asexual or selffertilizing species are effectively completely linked, in contrast to the substantial opportunities for recombination in outbreeding sexual species. Several key observations on the relation between patterns of evolution and levels of genetic recombination have been made. First, asexual or highly self-fertilizing species tend to be young in terms of the evolutionary timescale, suggesting that they become extinct more rapidly than their recombining relatives. Second, organelle and clonally transmitted genomes in several taxa show reduced levels of adaptation with respect to RNA and protein sequences. In Drosophila, regions where the recombination rate is low, such as near centromeres and telomeres, show reduced levels of codon bias — the non-random use of alternative codons encoding the same amino acid — suggesting a reduction in the ability of selection to maintain this aspect of molecular adaptation. Finally, in a number of different species, genomic regions with restricted recombination show lower levels of DNA sequence variation. One of the best examples is provided by non-recombining Y chromosomes, or neo-Y chromosomes formed by fusions between autosomes and sex chromosomes, which show degeneration of gene function and reduced genetic variability. Why should there be a relation between the amount of recombination and levels of variation and adaptation? As Fisher and Muller pointed out, the dynamics of a given gene are influenced by the evolutionary forces acting on the gene itself, as well as by forces acting at linked loci. Thus the predictions of single-locus population genetics must be modified when selection is acting on sets of linked loci. Here we describe some of the processes that may shape evolution when recombination is restricted over a large genomic region, and which may explain the above observations. These processes all reflect a general effect first quantified by Hill and Robertson: a locus linked to another locus under directional selection experiences a reduced effective population size, Ne. The extent of random fluctuations in allele frequencies due to finite population size, genetic drift, is inversely related to Ne; thus selective differences at one locus tend to enhance the effects of drift at a linked locus. According to the neutral theory of molecular evolution, mutation and genetic drift have a major influence on DNA sequence variation within species and on differences between species. Mutation creates new neutral variants, with no significant fitness effects, and genetic drift causes random changes in their frequencies until fixation or loss. Kimura showed that the level of variation within a population at a neutral locus is proportional to the product of Ne and the neutral mutation rate, μ, and that the rate of sequence evolution is equal to μ. A mutation with a selection coefficient, s, which measures the reduction or increase in the fitness of its carriers relative to that of the rest of the population, is effectively neutral if Ne s << 1. A strongly deleterious mutation, for which Ne s >> 1, will be rapidly eliminated, but a weakly deleterious mutation, with Ne s < 1, can persist and even become fixed in the population through genetic drift. Similarly a favourable mutation will have almost the same chance of loss from the population as a neutral mutation if Ne s < 1. Changes in Ne caused by different kinds of selection at linked loci can thus greatly affect both genetic variability and the efficacy of selection.