Résumé
Third generation of sequencing technologies yield long but very noisy reads. In this talk, i will present two approaches to correct error in long reads. First, a hybrid correction algorithm and tool, called LoRDEC, which is based on de Bruijn Graph. Then, a self correction method and tool, called LoRMA, that iteratively uses LoRDEC on the long reads to perform local correction, and then applies a multiple alignment procedure to further correct the reads. I will give an overview of the performances of both tools and comment on their ease of use.Work in collaboration with L. Salmela, R. Walve, E. Ukkonen from Helsinki University, and A. Zine El Aabidine from LIRMM.