iDINGO is a pathway-based method for estimating group-specific conditional dependencies and inferring differential networks between groups, based on genomic data. This can be done in a single-platform framework (for example, RNA-Seq data) or an integrative multi-platform framework (microRNA -> RNA -> Proteomics, where data from all three platforms are available for every sample).
We recommend filtering genomic data to fewer than 300 genes,
generally filtered using a pathway/pathways of interest. Single-platform
analyses are run using dingo
with an nxp matrix, where n is
the number of samples. Multi-platform analyses are run using
idingo
, with up to 3 separate data matrices containing the
same n samples. For both dingo
and idingo
, the
number of bootstraps is specified by B
(we recommend at
least 100). Parallel computing can speed this step up significantly, by
setting the number of cores
. Finally, the
plotNetwork
function plots the differential network
identified by dingo
or idingo
, based on a
user-specified p-value or differential score threshold.