Choosing Resources

Michael T. Zimmermann

2019-05-02

RITAN indexes multiple resources and choosing which of them are most relevant for your study can be a challenge. To help with this process, we provide below a set of examples from different types of studies and the thinking behind which resources were used.

Resources Indexed by default

library(RITANdata)
library(RITAN)
library(knitr)
kable( attr(network_list, 'network_data_sources') )
kable( sapply(geneset_list, length), col.names = c('# Genesets') )

While many resources contain information about protein complexes (obligate interactions) and protein-protein interactions (often transient), some use experimental techniques that are specific for physical ineractions. Determining whihc resources indicate physical, through-metabolite, and through-DNA (i.e. transcription factors) interactions, we recommend: 1) each resource’s primary publication 2) the pathguide website 3) the following guidelines

Example: Protein Complexes & Protein-Protein Interaction (PPI)

genes <- geneset_list$MSigDB_C5$APICAL_JUNCTION_COMPLEX
e <- network_overlap( genes, resources = c('CCSB','STRING'), minStringScore = 700 )
## We also strongly encourage use of the BioPlex database, which we do not distribute with RITAN in compliance with their licensing.

Example: Metabolic Network Interaction

genes2 <- geneset_list$MSigDB_C5$AMINE_METABOLIC_PROCESS
e2 <- network_overlap( genes, resources = c('PID','HumanNet') )

Example: Expression Signalling

genes2 <- geneset_list$MSigDB_C5$AMINE_METABOLIC_PROCESS
e2 <- network_overlap( genes, resources = c('ChEA','HumanNet') )