Genome-wide analyses of LINE-LINE-mediated nonallelic homologous recombination.

Michał Startek,Chad A. Shaw,Paweł Stankiewicz,Tomasz Gambin,Anna Gambin,Przemyslaw Szafranski,Ian M. Campbell,Patricia Hixson

doi:10.1093/nar/gku1394

Abstract

Nonallelic homologous recombination (NAHR), occurring between low-copy repeats (LCRs) >10 kb in size and sharing >97% DNA sequence identity, is responsible for the majority of recurrent genomic rearrangements in the human genome. Recent studies have shown that transposable elements (TEs) can also mediate recurrent deletions and translocations, indicating the features of substrates that mediate NAHR may be significantly less stringent than previously believed. Using >4 kb length and >95% sequence identity criteria, we analyzed of the genome-wide distribution of long interspersed element (LINE) retrotransposon and their potential to mediate NAHR. We identified 17 005 directly oriented LINE pairs located <10 Mbp from each other as potential NAHR substrates, placing 82.8% of the human genome at risk of LINE–LINE-mediated instability. Cross-referencing these regions with CNVs in the Baylor College of Medicine clinical chromosomal microarray database of 36 285 patients, we identified 516 CNVs potentially mediated by LINEs. Using long-range PCR of five different genomic regions in a total of 44 patients, we confirmed that the CNV breakpoints in each patient map within the LINE elements. To additionally assess the scale of LINE–LINE/NAHR phenomenon in the human genome, we tested DNA samples from six healthy individuals on a custom aCGH microarray targeting LINE elements predicted to mediate CNVs and identified 25 LINE–LINE rearrangements. Our data indicate that LINE–LINE-mediated NAHR is widespread and under-recognized, and is an important mechanism of structural rearrangement contributing to human genomic variability.

Highlights

Copy-number variation (CNV) contributes significantly both to human genetic variation as well as disease [1,2,3]
Rearrangements mediated by human endogenous retroviruses (HERVs) [9], a small subfamily of long retrotransposons comprising ∼0.8% of the human genome [9], suggest that the lower boundary on the length of the homologous region which is capable of mediating Nonallelic homologous recombination (NAHR) might be as low as few kb
Because of the relative abundance of transposons in the human genome compared to low-copy repeats (LCRs), they have the potential to mediate NAHR between a wider array of loci, potentially posing a significant contribution to genetic instability

Summary

Introduction

Copy-number variation (CNV) contributes significantly both to human genetic variation as well as disease [1,2,3]. In the vast majority of rearrangements characterized far, NAHR occurred between large segments of the human genome that are present in more than one copy known as low-copy repeats (LCRs or segmental duplications). These LCRs are typically >10 kb in size and share >97% DNA sequence identity [4,6,7,8]. Other mobile DNA elements [10] may be potential substrates for NAHR. If true, this would indicate that a signif-

Methods

Results

Conclusion