Run stabilized ICA - This module runs a stabilized version of Independent Component Analysis on a transcriptomic data set. The stability of the independent components is estimated through several runs of an ICA solver with random initializations. The results of the different runs are clustered and the centrotype components are considered to be the correct ones (i.e the stabilized ones).

GSEA analysis - This module applies Gene Set Enrichment Analysis to each stabilized ICA component (Subramanian et al., 2005). As input it uses the natural ranking given by the projection of each component in the "gene space" (i.e for each independent component, each gene is associated with a weight which reflects its importance in the definition of the component; this set of weights naturally ranks the genes for the given component).

ToppGene analysis - This module applies ToppGene functional enrichment analysis to each stabilized ICA component (Chen et al., 2009). As input it uses the "top contributing genes" of each component (i.e the genes associated with the highest weights in the gene space). The user can either fix the number of most weighted genes he wants to select as "top contributing genes", or set a threshold to select the most weighted genes.

OFTEN analysis - This module finds optimally functionnaly enriched subnetworks associated with the gene rankings provided by each stabilized ICA component. It uses PPI networks from the knowledge/undirected folder, and the methodology described in Kairov et al. 2O13.

Navicell visualization - This module maps the input transcriptomic data and/or the stabilized ICA components onto biological network maps, browsable with NaviCell tool. See this link for more details.

Metagene annotation - This module provides additional tools to annotate and interpret the stabilized ICA components via their projection in the "gene space" (i.e metagenes). It can correlate metagenes to certain gene properties (numerical or categorical) as well as to reference metagenes.

Metasample annotation - This module provides tools to annotate and interpret the stabilized ICA components via their projection in the "sample space" (i.e metasamples). It correlates the metasamples with additional sample annotations/features, using Spearman correlation for numerical features and Wilcoxon tests for categorical ones.

Meta-analysis with RBH graph - This module allows comparing all components extracted from different data sets, constructing a Reciprocally Best Hit (RBH) graph which represents reciprocal correlation relations between components extracted from different data sets.