PopAlu: population-scale detection of Alu polymorphisms

Abstract

Alu elements are sequences of approximately 300 basepairs that together comprise more than 10% of the human genome. Due to their recent origin in primate evolution some Alu elements are polymorphic in humans, present in some individuals while absent in others. We present PopAlu, a tool to detect polymorphic Alu elements on a population scale from paired-end sequencing data. PopAlu uses read pair distance and orientation as well as split reads to identify the location and precise breakpoints of polymorphic Alus. Genotype calling enables us to differentiate between homozygous and heterozygous carriers, making the output of PopAlu suitable for use in downstream analyses such as genome-wide association studies (GWAS). We show on a simulated dataset that PopAlu calls Alu elements inserted and deleted with respect to a reference genome with high accuracy and high precision. Our analysis of real data of a human trio from the 1000 Genomes Project confirms that PopAlu is able to produce highly accurate genotype calls. To our knowledge, PopAlu is the first tool that identifies polymorphic Alu elements from multiple individuals simultaneously, pinpoints the precise breakpoints and calls genotypes with high accuracy.

Keywords: Alu elements; Mobile element insertion; Paired-end sequencing; Polymorphism genotyping; Structural variation.

Figure 1. Example read alignments at an Alu deletion site.

Arrows show read directions. The blue part of the reads can be mapped to the reference outside of the Alu and the red part can be mapped to the Alu. (A) shows example reads from a haplotype that carries allele H₁. (B) shows example reads from a haplotype that carries allele H₀. A heterozygote diploid can have reads shown in both (A) and (B).

Figure 2. Example read alignment at an Alu insertion site.

Arrows show read directions. The blue part of the reads can be mapped to the reference and the red parts are clipped or mapped somewhere else in the reference. (A) shows example reads from a non-Alu haplotype. (B) shows example reads from an Alu insertion haplotype. A heterozygote diploid can have reads shown in both (A) and (B).

Figure 3. Example instance of our two-level voting system that determines the exact breakpoints of an Alu insertion.

At the first level, split-reads vote for a left and right breakpoint position within each individual. At the second level, individuals vote for the positions that received the largest numbers of votes at the first level to choose the final breakpoint positions AL and AR. In this example, position b is elected as AR and position e is elected as AL.