Here we present RAIDER, a tool for the de novo identification of elementary repeats. The problem of searching for genomic repeats without reference to a compiled profile library is important in the annotation of new genomes and the discovery of new repeat classes. Several tools have attempted to address the problem, but generally suffer either an inability to run at the whole-genome scale or loss of sensitivity due to sequence variation between repeat copies. To address this, Zheng and Lonardi define elementary repeats: building blocks that can be assembled into a repeat library, but allow for the filtering of spurious fragments. However, their tool was too slow for use on large input, and subsequent attempts to improve efficiency have been unable to deal with the expected variation between repeat instances. RAIDER addresses both these problems, implementing a novel algorithm for elementary repeat detection and incorporating the spaced seed strategy of PatternHunter to allow for copy variation. Able to process the human genome in under 6.4 hours, initial results indicate a coverage rate comparable to or better than that achieved by competing de novo search tool when paired with the library-based RepeatMasker. © 2013 Springer International Publishing.
CITATION STYLE
Figueroa, N., Liu, X., Wang, J., & Karro, J. (2013). RAIDER: Rapid Ab Initio Detection of Elementary Repeats. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8213 LNBI, pp. 170–180). https://doi.org/10.1007/978-3-319-02624-4_16
Mendeley helps you to discover research relevant for your work.