Dengue infection atlas - A systems biology approach using high-throughput data integration

Prasad Siddavatam, Purdue University

Abstract

Dengue virus (DENV) causes millions of infections around the world and is classified as an emerging pathogen in the USA. Though some of the steps in Dengue virus infection and replication are known, understanding of molecular mechanisms is still elusive. An unpublished microarray gene expression data from Dengue infected HuH7 liver cells at 18 hr and 24 hr p.i. was analyzed to identify differentially expressed genes (DEGs). Transcription of 30 genes was found to be perturbed (26 induced and four down-regulated). Pathway and functional analysis of the DEGs, strongly suggests that the Endoplasmic Reticulum plays an important role in Dengue virus infection. Integration of high-throughput data, DEGs, human proteins interacting with Dengue proteins (DHPINs) and human genes required for Dengue replication (DRNAi), confirms the significance of the Endoplasmic reticulum-associated protein-degradation pathway in DENV infection. DEGs identified in this work did not show extensive overlap with previously reported DEGs. Application of consistent analytical and statistical techniques to the gene expression data from ten DENV infection studies did not substantially improve the amount of overlap between the reported DEGs. Meta-analysis of the combined data using a leave-out dataset cross-validation approach, however, found a total of 1,568 differentially expressed genes. In analyses that include time points early in infection, pathways related to cancer and cell junction are enriched in DEGs, but in analyses in which early time-points are excluded, metabolic and cell signaling related pathways are enriched. This confirms that there is a progressive shift in the effect of the virus on cellular gene expression as infection progresses. Topological profiling, with respect to the human protein-interaction network (HPIN), of DEGs, DRNAi, and DHPINs from DENV shows that DEGs and DHPINs prefer proteins with high degree and betweenness centrality measures but DRNAi do not show a similar trend. However, DENV does not target proteins with high clustering coefficient. DHPINs do not interact with hub proteins (top 5% degree proteins) but rather interact with bottleneck proteins (top 5% betweenness centrality proteins). The combination of these studies substantially extends our understanding of the effects of Dengue virus on host molecular processes.

Degree

Ph.D.

Advisors

Gribskov, Purdue University.

Subject Area

Systematic|Bioinformatics

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS