The genomic landscape of short insertion and deletion polymorphisms in the chicken (Gallus gallus) Genome: a high frequency of deletions in tandem duplicates

Abstract

It is increasingly recognized that insertions and deletions (indels) are an important source of genetic as well as phenotypic divergence and diversity. We analyzed length polymorphisms identified through partial (0.25x) shotgun sequencing of three breeds of domestic chicken made by the International Chicken Polymorphism Map Consortium. A data set of 140,484 short indel polymorphisms in unique DNA was identified after filtering for microsatellite structures. There was a significant excess of tandem duplicates at indel sites, with deletions of a duplicate motif outnumbering the generation of duplicates through insertion. Indel density was lower in microchromosomes than in macrochromosomes, in the Z chromosome than in autosomes, and in 100 bp of upstream sequence, 5'-UTR, and first introns than in intergenic DNA and in other introns. Indel density was highly correlated with single nucleotide polymorphism (SNP) density. The mean density of indels in pairwise sequence comparisons was 1.9 x 10(-4) indel events/bp, approximately 5% the density of SNPs segregating in the chicken genome. The great majority of indels involved a limited number of nucleotides (median 1 bp), with A-rich motifs being overrepresented at indel sites. The overrepresentation of deletions at tandem duplicates indicates that replication slippage in duplicate sequences is a common mechanism behind indel mutation. The correlation between indel and SNP density indicates common effects of mutation and/or selection on the occurrence of indels and point mutations.

Figure 1.—

Size distribution of indels. (A) The size distribution of indels segregating in the chicken genome and (B) the size distribution of those observed in the chicken–turkey comparison.

Figure 2.—

Distribution of indel density in 1-Mb windows. Indel density was determined in 1-Mb windows across the chicken genome and normalized by shotgun read coverage. The smooth curve indicates the expected Poisson distribution.

Figure 3.—

Correlates of indel and SNP density in 1-Mb windows. (A) A significant correlation between indel density and SNP density. Both indel density (B) and SNP density (C) are correlated with GC level. However, correcting indel density and SNP density for GC does not remove the correlation (D), indicating that the same evolutionary forces act on indels and SNPs.

Figure 4.—

Indel density in different regions associated with protein-coding genes. Regions ar defined according to Ensembl annotations. Whiskers indicate the 95% confidence interval.

Figure 5.—

Size distributions of rooted insertions and deletions. Open bars represent insertions, while solid bars represent deletions.