The algorithms offered here allow an intuitive visualisation of multiple enriched functional annotation categories from DAVID found in hits from large-scale RNAi screens and can combine that view with information from functional genetic interactions from STRING.

One module (draw_david_network.m) visualises clusters of enriched DAVID functional annotation categories and their overlap as a network, in which nodes are individual clusters, and edges represent the fraction of genes shared between two nodes as a function of the cluster with fewest genes. All edges above 75% gene-identity (of the smaller cluster) were depicted. Edges of 33% gene-identity or higher are added for otherwise unconnected annotation cluster nodes. The code used to generate this visualization is implemented in Matlab, and requires Cytoscape for visualisation. This code has been tested for use on both PC and Mac.

A second module (draw_string_david_combination_network.m) combines the functional annotation clustering with the functional interactions described in STRING. To create this combined network view, the module assigns each gene to the cluster with the highest enrichment, in which the gene was present in the highest fraction of individual annotations. Genes not clustered in any of the DAVID clusters are left unconnected, and selected unclustered genes are placed back in the visualisation as inverted arrowheads. The module then adds all STRING interactions (yellow solid lines) with a confidence score of 0.9 or higher, and all interactions of 0.35 or higher between genes within the same functional annotation cluster. The code used to generate the visualisation is implemented in Matlab, and requires Cytoscape for visualization. This code has been tested for use on both PC and Mac.

Both modules contain extensive insightful comments within the code and come with a user manual.


Publications in which these methods were used:

  1. Mercer, J. et al. RNAi Screening Reveals Proteasome- and Cullin3-Dependent Stages in Vaccinia Virus Infection. Cell Reports 2, 1036–1047 (2012)
  2. Snijder, B. et al. Single-cell analysis of population context advances RNAi screening at multiple levels. Molecular Systems Biology 8, (2012)