Fig. 4From: Reference-free phylogeny from sequencing dataThe main idea of the efficiency improvements in Efficiency improvements section. Instead of pairwise alignment of all reads and contigs vs. all reads and contigs, we run the exact quadratic alignment only for a small subset of those. This subset of candidates is identified based on the q-gram distance, which runs in linear time. See Fig. 1 for meaning of the caption numbersBack to article page