DYNAMO

The study of function and organization of large-scale cellular processes can be aided by integrative approaches that combine gene expression data analysis and molecular network representation of biological processes. In this study, we compare the temporal evolution of the global gene response of a key immune cell, the human dendritic cell, to infection by influenza virus strains that vary in their human infectivity and pathogenicity. To achieve this, we implemented a method (DYNAMO) that allows the comparison of activation patterns of functionally related groups of genes (modules) with respect to time. Using an immune-specfic network inferred from a comprehensive large-scale public data compendium, DYNAMO identfies modules comprised of regulated, functionally connected nodes. Rather than identifying modules as a group of genes showing correlated regulation, DYNAMO defines modules as functionally related groups of genes that show similar patterns of regulation for different viral infections. Importantly, DYNAMO uses time-course data for module discovery and comparison of the temporal dynamics of module behavior between different viruses. The incorporation of temporal dynamics is critical the identification of subnetworks showing conserved and differential behavior in response to different viruses, and enables DYNAMO to provide mechanistic insight into the host response to these different influenza strains.

Contact: Dr. Elena Zaslavsky

RESULTS

Each pair of immune responses to influenza strains (Cal, Tx, Brevig, NC) was put through DYNAMO to find either time-shifted conserved or differential subnetworks. The tables with links to the results for each type of search are shown below. For each comparison, the link to its random/real curve is found at the top of the listed results. These curves help assess the significance of the found subnetworks.

Results for pairwise searches for temporally-shifted conserved subnetworks:

DYNAMO finds conserved subnetworks whose optimal similarity arises when one response is shifted relative to the other in time. The considered time shifts are up to 80 min in either direction, done in 20 min increments.

Table 1. Pairwise searches for temporally-shifted conserved subnetworks
Brevig Cal NC Tx
Brevig N/A Discovered subnetworks Discovered subnetworks Discovered subnetworks
Cal N/A Discovered subnetworks Discovered subnetworks
NC N/A Discovered subnetworks
Tx N/A

Results for pairwise searches for differential subnetworks:

This differential search looked for subnetworks showing opposing patterns of activity.

Table 2. Pairwise searches for differential subnetworks
Brevig Cal NC Tx
Brevig N/A Discovered subnetworks Discovered subnetworks Discovered subnetworks
Cal N/A Discovered subnetworks Discovered subnetworks
NC N/A Discovered subnetworks
Tx N/A