Schematic of JointCluster algorithm with scaling heuristic.
Clustering tree produced from a simultaneous clustering of two networks
. JointCluster can handle any number of networks, and the scaling heuristic transitions from a sum graph
to the smallest of the sparse cuts in the individual graphs
in increments of a scale parameter. At scale
, the graphs being analysed are
. Only some steps of the scaling heuristic are shown here for simplicity, and complete description is provided in Methods
. Graph-specific clusters are shown as different shaded regions within a larger set of nodes.
Articles from PLoS Computational Biology are provided here courtesy of
Public Library of Science