Proposal of a New Method for de Novo DNA Sequence Assembly Using de Bruijn Graphs

In this work, we propose a method where all main concerns in the construction of a DNA sequence assembler is encompassed in a single computational approach. The graph that represents sequences and their relationships is broken into simpler components through a matching approach on bipartite graphs, which we show to be very suitable for parallelism. Next, a simple heuristic is used to combine the components into long paths in the graph, which correspond to contigs in the genome. Experiments with five datasets have shown high quality assemblies.