P0998 HelitronScanner: A General Helitron Detection Tool

Chunguang Du , Montclair State University, Montclair, NJ
Wenwei Xiong , Montclair State University, Montclair, NJ
Jason Caronna , Montclair State University, Montclair, NJ
Limei He , Waksman Institute, Rutgers University, Piscataway, NJ
Hugo Dooner , Waksman Institute, Rutgers University, Piscataway, NJ
As a new class of transposable elements, Helitrons feature the remarkable ability to capture gene sequences, which makes them of considerable potential evolutionary importance. Unlike other DNA transposable elements, Helitrons possess low conservation in terms of sequence or structural features. Helitrons have been mainly reported to exist in plant genomes. So far, there is no effective way to detect Helitrons in all species. Here we propose a Local Combinational Variable (LCV) approach for developing the HelitronScanner program, aiming to scan whole genomes of all available species and find Helitrons within them. First, LCVs are generated from high quality Helitron sequences from the maize genome. These LCVs, as well as some biological characteristics of Helitrons, are employed to develop the HelitronScanner program. Pair ends are adopted in our program, where a possible Helitron end should have a matching score above the empirical threshold while also meeting the biological requirements. To validate prediction quality, we compared detected Helitrons in the maize genome by both HelitronFinder (a previous version of our Helitron detection tool) and HelitronScanner. The comparison is conducted by matching flanking sequences (combining together 100 bps from both flanks) of putative Helitrons against maize genome data of 6 inbred lines from a collaborator’s lab. The comparison feedback is then used to improve HelitronScanner. Analysis of putative Helitrons from other available species (i.e. sorghum, poplar, rice, Arobadopsis, soybean) is also expected to boost the performance of HelitronScanner.