A Tale of Optimizing the Space Taken by de Bruijn Graphs

In the last decade in bioinformatics, many computational works have studied a graph data structure used to represent genomic data, the de Bruijn graph. It is closely tied to the problem of genome assembly, i.e. the reconstruction of an organism's chromosomes using a large collection of overlapping short fragments. We start by highlighting this connection, noting that assembling genomes is a computationnally intensive task, and then focus our attention on the reduction of the space taken by de Bruijn graph data structures. This extended abstract is a retrospective centered around my own previous work in this area. It complements a recent review [10] by providing a less technical and more introductory exposition of a selection of concepts.