We provide the efficient implementations of the algorithmic steps (i.e., Steps 3 and 4). Both source codes and README files can be downloaded below.
Step 3: Tensor-based frequent dense subgraph identification algorithm. This software discovers frequent dense subgraphs from multiple unweighted biological networks. By modeling multiple networks as a tensor, we formulate the problem of discovering frequent patterns as an optimization problem with sparse constraint and employ the multi-stage convex relaxation method. It can find frequent patterns across a large collection of massive unweighted networks.
Tensor-based frequent dense subgraph algorithm manual
Click here to download source code.
Step 4: A counting algorithm for final pattern recovery. This software recovers a frequent dense subgraph/module in the original chromatin interaction graphs from the contracted subgraphs obtained in Step 3.
Click here to download source code.
