Skip to main content
Fig. 3 | BMC Genomics

Fig. 3

From: Graph mining for next generation sequencing: leveraging the assembly graph for biological insights

Fig. 3

Multilevel graph set and hybrid graph. a Multilevel graph set. G0 is the most granular graph created from all of the read overlap relationships generated during read alignment. Each read is assigned to a node in G0 and overlap relationships are assigned to edges. Weights on edges reflect the length of the overlap relationship. Heavy edge matching and node merging is applied to create a spectrum of graphs. Clusters of reads are formed as nodes are merged at each graph level. b Hybrid graph. Starting with the simplest graph, in this case G3, Focus attempts to assemble the read clusters represented by each node. If the reads assemble into a single contig, then their corresponding node is selected as that cluster’s best representative. All nodes that are selected by Focus in G3 as well as nodes not selected are used to create G 3 of the hybrid graph spectrum. If a node is not selected in G3, then its children nodes in the next graph level, in this case G2, will be evaluated. The graph G 2 will be created from the nodes selected from G3 and G2 as well as from the nodes that were not selected in G2.. We denote the final graph G 0 as the hybrid graph as it will contain all of the best representatives from Gn … G0. In (b), graph level integration of G3 … G1 better represents a split in the overlap graph

Back to article page