Preparation of Non-overlapping Transposable Elements (TEs) Annotation by Interval Tree.

Abstract

Transposable elements (TEs) are a major source of PIWI-interacting RNAs (piRNAs), therefore properly assigning piRNA library sequencing reads to the TEs from which they were derived is important for accurate assessment of piRNA biology. When calculating the abundance of small RNA-seq reads mapping to various TEs, a non-overlapping TE annotation is preferable because reads mapping to more than one genomic feature will often be excluded when counting reads. However, most unmodified TE annotations contain some degree of overlap between TE features. Here, I outline the principle and provide all scripts needed to resolve such overlapping regions of TE annotations to a single best TE annotation leveraging a computationally efficient tree algorithm. Non-overlapping annotations generated by this method can be directly used in commonly used read counting software.