Section: New Results

High-throughtput sequence processing

• We published a book chapter on bioinformatics algorithms for GPU/manycore processors [16] .

• Within the PhD of T. T. Tran, we proposed a new bit-parallel algorithm, extension of [39] , as well as a new indexing structure adpated to GPUs [13] .

• We proposed a new index structure specifically designed for reads produced by high throughput sequencers. It can deal both with variable or fixed length reads and can index reads in usually less memory than other classical solutions. This index has been implemented and is available online [10] .

• We characterised the number of elements to be reordered when updating a full-text index such as a suffix array. We finally concluded that this number tends to be poly-logarithmic in the input length for DNA sequences [5] .