LCR-Finder: A de novo low copy repeat finder for human genome

Abstract

Low copy repeats (LCRs) are reported to trigger and mediate genomic rearrangements and may result in genetic diseases. The detection of LCRs provides help to interrogate the mechanism of genetic diseases. The complex structures of LCRs render existing genomic structural variation (SV) detection and segmental duplication (SD) tools hard to predict LCR copies in full length especially those LCRs with complex SVs involved or in large scale. We developed a de novo computational tool LCR-Finder that can predict large scale (>100Kb) complex LCRs in a human genome. Technical speaking, by exploiting fast read alignment tools, LCR-Finder first generates overlapping reads from the given genome, aligns reads back to the genome to identify potential repeat regions based on multiple mapping locations. By clustering and extending these regions, we predict potential complex LCRs. We evaluated LCR-Finder on human chromosomes, we are able to identify 4 known disease related LCRs, and predict a few more possible novel LCRs. We also showed that existing tools designed for finding repeats in a genome, such RepeatScout and WindowMasker are not able to identify LCRs and tools designed for detecting SDs also cannot report large scale full length complex LCRs. © 2013 Springer-Verlag.