Homepage Inria website

Section: New Results

High-throughtput sequence processing

  • Analysis of immunological rearrangements for leukemia diagnosis and monitoring. High-throughput sequencing is spreading in the hospitals and many classical routines are now being transferred to this new technology. However in the specific case of lymphocyte monitoring, some complications arise. Classical bioinformatics software tools do not apply to the specificity of lymphocyte rearrangements. That is why we developed the software Vidjil (see 5.2 ) together with Lille hospital. This work has been published [5] and was also presented, as a poster, during the annual conference of the American Society of Hematology (ASH) [13] . We are now members of the EuroClonality-NGS work group which aims at providing a standardized way of monitoring leukemia using high-throughput sequencing at the European level.

  • New seeds for approximate pattern matching. We addressed the problem of approximate pattern matching using the Levenshtein distance. Given a text T and a pattern P, find all locations in T that differ by at most k errors from P. For that purpose, we proposed a filtration algorithm that is based on a novel type of seeds, combining exact parts and parts with a fixed number of errors, that we called 01*0 seed. Implementation has been performed on a Burrows-Wheeler transform. Experimental tests show that the method is specifically well-suited to search for short patterns (<50 letters) on a small alphabet (e.g. DNA alphabet) with a medium to high error-rate (7 %–15 %). This work has been published in [9] , and has a large number of applications in computational biology, such as finding microRNA targets, for example.

  • Spaced seeds and Transition seeds. This year, two collaborative works have been published on the topic of spaced seeds and derivated models. The first work, resulting from a collaboration with Martin C. Frith from the Computational Biology Research Center (Tokyo), increases the sensivity of several search tools (among them, LAST, LASTZ, YASS,...) by computing specific seeds adapted to transition ratios observed during Eucaryotic comparisons. This work has been published in [3] , together with the design of seeds obtained. The second work, issued from collaboration with Donald E.K. Martin from the Department of Statistics of the North Carolina State University (Raleigh), deals with the coverage of spaced seeds and shows how this criterion helps selecting good seeds for SVM string-kernels and alignment-free distances. This work has been published in [6] .