The distribution of deleterious genetic variation in human populations
Kirk E Lohmueller
Population genetic studies suggest that most amino-acid changing mutations are deleterious. Such mutations are of tremendous interest in human population genetics as they are important for the evolutionary process and may contribute risk to common disease. Genomic studies over the past 5 years have documented differences across populations in the number of heterozygous deleterious genotypes, numbers of homozygous derived deleterious genotypes, number of deleterious segregating sites and proportion of sites that are potentially deleterious. These differences have been attributed to population history affecting the ability of natural selection to remove deleterious variants from the population. However, recent studies have suggested that the genetic load may not differ across populations, and that the efficacy of natural selection has not differed across human populations. Here I show that these observations are not incompatible with each other and that the apparent differences are due to examining different features of the genetic data and differing definitions of terms.
Another relevant reference is Zuk et al. 2013 (http://www.pnas.org/content/early/2014/01/16/1322563111). They also come to the conclusion that demography has little effect on the combined frequency of deleterious alleles in a population (though does have an effect on the distribution of individual allele frequencies).