Section: New Results

High Throughput Sequence Analysis

Participants : Rayan Chikhi, Erwan Drezen, Dominique Lavenier, Claire Lemaitre, Nicolas Maillet, Pierre Peterlongo.

• Comparing metagenomes. This research aims to define new ways of comparing billions of sequences generated by NGS sequencers. Standard techniques don’t scale with such volume of data, both in terms of memory fingerprint and execution time. We have successfully tested a new method based on probabilistic data structures (Bloom filter) allowing large sets of sequences to be indexed in a short time on standard computers. [25]

• Bank-to-bank comparison. In cooperation with the Korilog company we improve the PLAST technology developed for bank-to-bank sequence similarity search. Structuration of the index has been revisited to reduce the memory fingerprint and the execution time. The Korilog company has successfully integrated this improvments sofware component in its own software and has just began its promotion with promising responses from several potential clients.[Korilog promotion]