Abstract

Next-generation sequencing technologies are currently bringing about a revolution in the way genomic sequencing is used in biology. Thanks to this tremendous reduction in costs, many DNA -based assays are now within reach of even modest research groups, and new ones are being quickly developed.

However, these technologies have the drawback of reading only very short fragments. These ‘reads’ can easily be mapped onto a known reference genome, but are much more difficult to handle in the case of unstudied species.

I will be presenting Velvet, a software package which allows users to reconstruct, or assemble, new genomes using these very short reads only. I will describe the theory behind Velvet, and especially its data structure, the de Bruijn graph. This kind of graph has is not only very useful in genome assembly, but also in analysing large scale rearrangements in the genome.