Biobank-scale inference of multi-individual identity by descent and gene conversion

Abstract

We present a method for efficiently identifying clusters of identical-by-descent haplotypes in biobank-scale sequence data. Our multi-individual approach enables much more computationally efficient inference of identity by descent (IBD) than approaches that infer pairwise IBD segments and provides locus-specific IBD clusters rather than IBD segments. Our method's computation time, memory requirements, and output size scale linearly with the number of individuals in the dataset. We also present a method for using multi-individual IBD to detect alleles changed by gene conversion. Application of our methods to the autosomal sequence data for 125,361 White British individuals in the UK Biobank detects more than 9 million converted alleles. This is 2,900 times more alleles changed by gene conversion than were detected in a previous analysis of familial data. We estimate that more than 250,000 sequenced probands and a much larger number of additional genomes from multi-generational family members would be required to find a similar number of alleles changed by gene conversion using a family-based approach. Our IBD clustering method is implemented in the open-source ibd-cluster software package.

Figure 1

Transitivity of IBD (A) The coalescent tree relationship at a given point in the genome is shown for haplotypes $h_1,h_2\text{,}$ and $h_3$, which are mutually IBD at this location. Haplotypes $h_1$ and $h_2$ have common ancestor X, while haplotypes $h_1$ and $h_3$, and haplotypes $h_2$ and $h_3\text{,}$ have common ancestor Y. (B) Pairwise IBD status (black for IBD, white for non-IBD) is shown for the three pairs of haplotypes along a region of the chromosome around the focal position (denoted ^∗). The IBD extends to either side of the focal point until reaching a point of recombination on one of the ancestral lineages. Although the IBD sharing between haplotypes $h_1$ and $h_2$, and between haplotypes $h_2$ and $h_3$, is long and may exceed a pre-defined length threshold, the IBD between haplotypes $h_1$ and $h_3$ is relatively short and may not meet the length threshold for pairwise IBD sharing.

Figure 2

IBD transitivity with and without trimming IBS segments IBD and IBS in a genomic region is shown for the three pairings of three haplotypes (haplotypes $h_1,h_2\text{,}$ and $h_3$). (A) The IBD between haplotypes $h_1$ and $h_2$ is derived from a different recent common ancestor than that of the IBD between haplotypes $h_2$ and $h_3$. IBS that is not due to the recent common ancestors is incorrectly called as IBD at the ends of the IBD segments. As a result, transitivity leads to a region of IBS being incorrectly called as IBD between haplotypes $h_1$ and $h_3$. (B) A trim is applied to the ends of the pairwise IBS regions, and no IBD is called between haplotypes $h_1$ and $h_3$.

Figure 3

Observed length of inferred gene conversion tracts Observed length of inferred gene conversion tracts in (A) simulated data (125,000 individuals with 20 regions of length 10 Mb) and (B) UK Biobank White British. The observed length of an inferred tract is the distance between the first changed allele and the final changed allele (inclusive), which will generally be significantly shorter than the actual length of the underlying gene conversion tract. Only lengths >1 bp are shown: 80.6% of observed lengths were 1 bp in the large simulate data, and 82.9% of observed lengths were 1 bp in the UK Biobank White British data.

Figure 4

IBD cluster sizes in the UK Biobank White British autosomal sequence data Cluster size is shown on the x axis for cluster sizes of $\le 3$ in the left panel and $\ge 3$ in the right panel. The y axis shows the proportion of haplotypes that are in IBD clusters having that size.