New Results About the Linearization of Scaffolds Sharing Repeated Contigs

Abstract

Solutions to genome scaffolding problems can be represented as paths and cycles in a "solution graph". However, when working with repetitions, such solution graphs may contain branchings and, thus, they may not be uniquely convertible into sequences. Having introduced various ways of extracting the unique parts of such solutions, we extend previously known NP-hardness results to the case that the solution graph is planar, bipartite, and subcubic, and show that there is no PTAS in this case.

Linearizing Genomes: Exact Methods and Local Search

“…In previous work, we show that the variants of Semi-Brutal Cut according to all presented scoring functions are NP-complete [15]. In [10], we explore special classes of graphs, namely bipartite, planar with bounded degree, analyzing complexity and approximability, showing that even in very restricted cases, the problem is hard to solve. We also proposed a 2-approximation algorithm under the weight score and a 4-approximation under the cut score.…”