Human Retrotransposons and Effective Computational Detection Methods for Next-Generation Sequencing Data

Abstract

Transposable elements (TEs) are classified into two classes according to their mobilization mechanism. Compared to DNA transposons that move by the "cut and paste" mechanism, retrotransposons mobilize via the "copy and paste" method. They have been an essential research topic because some of the active elements, such as Long interspersed element 1 (LINE-1), Alu, and SVA elements, have contributed to the genetic diversity of primates beyond humans. In addition, they can cause genetic disorders by altering gene expression and generating structural variations (SVs). The development and rapid technological advances in next-generation sequencing (NGS) have led to new perspectives on detecting retrotransposon-mediated SVs, especially insertions. Moreover, various computational methods have been developed based on NGS data to precisely detect the insertions and deletions in the human genome. Therefore, this review discusses details about the recently studied and utilized NGS technologies and the effective computational approaches for discovering retrotransposons through it. The final part covers a diverse range of computational methods for detecting retrotransposon insertions with human NGS data. This review will give researchers insights into understanding the TEs and how to investigate them and find connections with research interests.

Keywords: computational tools; next-generation sequencing (NGS); retrotransposons; transposable elements.

Figure 1

Composition of non-LTR retrotransposons in the human. Non-LTR retrotransposons constitute approximately 34% of the entire human genome. A total of 17% of L1 elements, 11% of Alu elements, and 0.2% of SVAs belong to non-LTR retrotransposons. As some of the elements are still active in humans, they cause genetic disorders and contribute to diversity.

Figure 2

Structure of L1, Alu, and SVA elements. (a) L1 elements have ~6 kb length. ORF1 in L1 element encodes RNA-binding protein. ORF2 encodes endonuclease and reverse transcriptase for self-mobilization. (b) Alu elements have a dimeric structure with a length of 300 bp. The two monomers are present on both sides of the A-rich region. (c) The canonical length of SVA elements is 2 kb. They contain five distinctive regions.

Figure 3

Illustration of discordant read, read-pair and soft-clipped reads. (a) A discordant read indicates that one-end read is fully mapped to the reference, but another end is not mapped to the reference. (b) A read pair provides putative insertion site and gives information about breakpoint interval based on 5′cluster at one end and 3′cluster at the other end. (c) A soft-clipped read (split read) refers a mate pair where one part is partially mapped to the reference. Hence, the truncated read contains both reference and novel insertion sequences.