CanMod2

CanMod2 is a computational pipeline to identify gene regulatory modules containing 6 main steps:

• Step 1: Cluster the differentially expressed (DE) genes based on GO-term similarity and generate Gene Clusters (GCs)
• Step 2: Identify regulators (i.e., miRNAs or transcription factors) for each DE gene
• Step 3: Cluster the regulators based on shared target similarity and generate Regulator Clusters (RCs)
• Step 4: Generate modules using the regulators and the target GCs

Iteratively refining modules running following two steps:

• Step 5: Refine each module using biclustering
• Step 6: Merge modules that share high similarity

Run CanMod2 with the command Rscript CANMOD_v2.R for sample run, or Rscript CANMOD_v2.R *data_name* where data_name.rda exists under data folder with the following variables:

• mRNA: A data frame containing gene expression data (Rows are all DE genes and columns are samples).
• miRNA: A data frame containing microRNA expression data (Rows are all DE microRNAs and columns are samples).
• methyl: A data frame containing DNA methylation data (Rows are all DE genes and columns are samples).
• cnv: A data frame containing copy number aberration data (Rows are all DE genes and columns are samples). ! Samples in these four variables will be in the same order.

---

User Options

• Hyperparameters (lines 34-36):
  • drop.thr: Stopping criteria for iterating Step 5 and 6. (default is 0.05)
  • GC_thr: Gene cluster (GC) threshold (default is 0.45)
  • RC_thr: Regulator cluster (RC) threshold (default is 0.10)
• To input user-defined GCs instead of CANMOD-generated GCs, provide the following files under main folder:
  • sample_user.defined.gene.cluster.list.rda: File having one list named gene.cluster.list containing the list of GCs
  • sample_user.defined.similarities.rda: File having one matrix named de.gene.bp.sim containing the similarity between genes in GCs