Characterization of full-length LINE-1 insertions in 154 genomes

Abstract

Long interspersed nuclear elements (LINEs) are retrotransposons that contribute to genetic variation in the human genome. LINE-1 elements in larger-scale studies are challenging to identify using sequencing technologies due to cost and scalability. We developed an approach using optical mapping for detection of full-length LINE-1 insertions and 10× sequencing for confirmation. We found 51 true positive full-length LINE-1 insertions, of which 4 are novel insertions, in NA12878. Repeating our analysis on a larger sample set representing 26 populations, we identified 329 full-length LINE-1 elements, of which 123 are novel. 24.8% of these 329 LINE-1 insertions were shared amongst all 5 superpopulations (AFR, AMR, EUR, EAS, SAS). The African superpopulation has a higher percentage of population-specific LINE-1 insertions than any other superpopulation. These data indicate that our approach can provide high-speed, cost-effective, and increased accuracy for LINE-1 detection. These data also provide an insight into variations of LINE-1 elements between different populations.

Keywords: Genomics; LINE-1.

Figure 1.

Schematic of LINE-1 detection and confirmation. (A) Label sites on LINE-1 references and their positions within a 6 kb insertion. The yellow dots represent Nt.BspQ1 motif sites and orange dots represent DLE-1 motif sites. Each value on the y-axis represents one out of 300 references, and the x-axis is the position from 0-6 kb within LINE-1. (B) DLE-1 motif positions and Nt.BsqQ1 motif positions are combined to determine the distances between DLE-1 - Nt.BspQ1 and Nt.BspQ1 - DLE-1. (C) Sequencing confirmation of LINE-1 insertion using 10x data shows soft clipped reads (orange) align to LINE-1 but not to hg38 (green).

Figure 2.

Detection of LINE-1 insertion in optical mapping data. (A) Optical mapping visualization of sample HG04006 in region chr6:72082349-72108286. Green bars represent hg38 reference and blue bars represent sample contig. Black nicks in both DLE-1 and Nt.BspQ1 datasets align to hg38. Orange nicks in the DLE-1 contig and yellow nicks in the Nt.BsqQ1 contig are nicks that do not align to any sites in hg38. The positions of unaligned DLE-1 and Nt.BsqQ1 nick sites within an insertion are used for calculating distances. (B) Soft clipped reads are aligned to LINE-1 reference at 0 bp and 6 kb.

Figure 3.

Population distribution of all LINE-1 insertion loci. (A) Distribution of LINE-1 insertions between five superpopulations: Loci found only in African (red), East Asian (green), South Asian (orange), European (purple), and American (blue) populations. Percentages are also calculated for LINE-1 insertions shared amongst 2-4 superpopulations (gray) and all populations (black). (B) PCA plot of LINE-1 insertions for all five superpopulations.