A genomewide assessment of inbreeding depression: gene number, function, and mode of action.

Although the genetic basis of inbreeding depression is still being debated, most fitness effects are thought to be the result of increased homozygosity for recessive or partially recessive deleterious alleles rather than the loss of overdominant genes. It is unknown how many loci are associated with inbreeding depression, the genes or gene pathways involved, or their mode of action. To uncover genes associated with variation in fitness following inbreeding, we generated a set of inbred lines of Drosophila melanogaster for which only the third chromosome varied among lines and measured male competitive reproductive success among these lines to estimate inbreeding depression. Male competitive reproductive success for different lines validated our prediction that equally inbred lines show variation in inbreeding depression. To begin to assess the molecular basis of inbreeding depression for male competitive reproductive success, we detected variation in whole-genome gene expression across these inbred lines with commercially available high-density oligonucleotide microarrays. A total of 567 genes were differentially expressed among these inbred lines, indicating that inbreeding directly or indirectly affects a large number of genes: genes that are disproportionately involved in metabolism, stress and defense responses. Subsequently, we generated a set of outbred lines by crossing the highest inbreeding depression lines to each other and contrasted gene expression between parental inbred lines and F(1) hybrids with transcript abundance as a quantitative phenotype to determine the mode of action of the genes associated with inbreeding depression. Although our results indicated that approximately 75% of all genes involved in inbreeding depression were additive, partially additive, or dominant, about 25% of all genes expressed patterns of overdominance. These results should be viewed with caution given that they may be confounded by issues of statistical inference or associative overdominance.