Integrated Mobile Element Scanning (ME-Scan) method for identifying multiple types of polymorphic mobile element insertions

Abstract

Background: Mobile elements are ubiquitous components of mammalian genomes and constitute more than half of the human genome. Polymorphic mobile element insertions (pMEIs) are a major source of human genomic variation and are gaining research interest because of their involvement in gene expression regulation, genome integrity, and disease.

Results: Building on our previous Mobile Element Scanning (ME-Scan) protocols, we developed an integrated ME-Scan protocol to identify three major active families of human mobile elements, AluYb, L1HS, and SVA. This approach selectively amplifies insertion sites of currently active retrotransposons for Illumina sequencing. By pooling the libraries together, we can identify pMEIs from all three mobile element families in one sequencing run. To demonstrate the utility of the new ME-Scan protocol, we sequenced 12 human parent-offspring trios. Our results showed high sensitivity (> 90%) and accuracy (> 95%) of the protocol for identifying pMEIs in the human genome. In addition, we also tested the feasibility of identifying somatic insertions using the protocol.

Conclusions: The integrated ME-Scan protocol is a cost-effective way to identify novel pMEIs in the human genome. In addition, by developing the protocol to detect three mobile element families, we demonstrate the flexibility of the ME-Scan protocol. We present instructions for the library design, a sequencing protocol, and a computational pipeline for downstream analyses as a complete framework that will allow researchers to easily adapt the ME-Scan protocol to their own projects in other genomes.

Keywords: AluYb; High-throughput sequencing; LINE-1; ME-Scan; Mobile element insertion; Retrotransposon; SVA.

Fig. 1

ME-specific amplification during ME-Scan library construction. For each ME type library, two rounds of nested amplification are performed. The ME-specific amplification primers (ME1 and ME2) are shown as thin arrows above the ME consensus and the amplification directions are indicated by the arrows. First-round amplification primers (ME1) are biotinylated (green star) for enrichment, and the second-round nested primers (ME2) include the Illumina sequencing adaptor (orange box). Different components of AluYb, SVA, and L1HS consensuses are labelled. The final paired-end sequencing reads from the resulting sequencing libraries are represented with blue arrows (ME Reads) and black arrows (Flanking Reads), respectively. Blue box: ME sequence; grey box: flanking genomic region; green star: biotin; orange box: Illumina sequencing adaptor

Fig. 2

Computational data analysis overview. a) The paired-end sequencing reads. Sequencing reads from the pooled libraries are represented by red (ME Reads) and blue arrows (Flanking Reads), respectively. b) Read filtering. The ME Reads are compared to the targeted ME consensus to identify recent insertions and are filtered based on the BLAST bit-score cutoff. The Flanking Reads are mapped to the reference genome and are filtered based on the mapping quality score cutoff. c) Flanking Read clustering and insertion loci identification. Filtered Flanking reads that are within a 500 bp sliding window are clustered into a candidate insertion locus and the genomic position closest to the ME Read is selected as the insertion position (marked with a star). Black box: clustering window

Fig. 3

Sensitivity analysis for determining proper TPM and UR cutoffs. Using presumably fixed reference MEIs as true positives, the sensitivity is calculated under different TPM and UR cutoffs for AluYb, L1HS, and SVA candidate loci, respectively. The average individual sensitivity (left panel) and overall sensitivity (right panel) for the 36 YRI samples are shown. The sensitivity is shown as the percentage of presumably fixed insertions being identified for each cutoff. The heatmap color corresponds to the sensitivity, as indicated in the color bar on the right of each plot