Fig. 3From: A graph-based approach for the visualisation and analysis of bacterial pangenomesGene association networks provide relationship between gene presence and host genome characteristics. A A filtered gene-gene association graph visualised in 2D space. Each node represents a gene and weighted edges correspond to the Jaccard similarity coefficient between each pair of genes (JSC > 0.550). High and low frequency genes have been removed (n < 7; n > 770), as have smaller components (n < 10), for visual clarity. Nodes are coloured by cluster (MCLi = 1.50). B Four gene distribution profile plots, representing commonly observed distributions: Cluster 1, near-core genes; Cluster 3, set of genes showing a near lineage-specific distribution (CC45); Cluster 25, host-associated (aves) set of genes; Cluster 28, widely distributed but not lineage-specific set of genes. Each bar in a plot represents an individual genome, with bar height equal to the proportion of genes in the cluster present in that genome. Coloured horizontal bars below the x-axis represent the host (upper) and clonal complex (lower)Back to article page