Limits...
Nodes with high centrality in protein interaction networks are responsible for driving signaling pathways in diabetic nephropathy.

Abedi M, Gheisari Y - PeerJ (2015)

Bottom Line: We found 49 genes to be variably expressed between the two groups.Interestingly, the central nodes were more informative for pathway enrichment analysis compared to all network nodes and also 49 differentially expressed genes.In conclusion, we here show that central nodes in protein interaction networks tend to be present in pathways that co-occur in a biological state.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Genetics and Molecular Biology, Isfahan University of Medical Sciences , Isfahan , Iran.

ABSTRACT
In spite of huge efforts, chronic diseases remain an unresolved problem in medicine. Systems biology could assist to develop more efficient therapies through providing quantitative holistic sights to these complex disorders. In this study, we have re-analyzed a microarray dataset to identify critical signaling pathways related to diabetic nephropathy. GSE1009 dataset was downloaded from Gene Expression Omnibus database and the gene expression profile of glomeruli from diabetic nephropathy patients and those from healthy individuals were compared. The protein-protein interaction network for differentially expressed genes was constructed and enriched. In addition, topology of the network was analyzed to identify the genes with high centrality parameters and then pathway enrichment analysis was performed. We found 49 genes to be variably expressed between the two groups. The network of these genes had few interactions so it was enriched and a network with 137 nodes was constructed. Based on different parameters, 34 nodes were considered to have high centrality in this network. Pathway enrichment analysis with these central genes identified 62 inter-connected signaling pathways related to diabetic nephropathy. Interestingly, the central nodes were more informative for pathway enrichment analysis compared to all network nodes and also 49 differentially expressed genes. In conclusion, we here show that central nodes in protein interaction networks tend to be present in pathways that co-occur in a biological state. Also, this study suggests a computational method for inferring underlying mechanisms of complex disorders from raw high-throughput data.

No MeSH data available.


Related in: MedlinePlus

Enrichment of the PPI network and selection of central nodes for pathway enrichment analysis can determine pathways essentially related to DN.The 49-node network was extended with maximum two interactive genes for each node. The initial nodes selected from the microarray experiment are depicted with red color and enriched nodes with black (A). In this expanded network, 34 genes were selected as nodes with high centrality. Pathway enrichment analysis with these “central genes” disclosed 62 highly connected pathways related to DN. Pathways with adjusted p-value <0.05 are shown (B).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4636410&req=5

fig-3: Enrichment of the PPI network and selection of central nodes for pathway enrichment analysis can determine pathways essentially related to DN.The 49-node network was extended with maximum two interactive genes for each node. The initial nodes selected from the microarray experiment are depicted with red color and enriched nodes with black (A). In this expanded network, 34 genes were selected as nodes with high centrality. Pathway enrichment analysis with these “central genes” disclosed 62 highly connected pathways related to DN. Pathways with adjusted p-value <0.05 are shown (B).

Mentions: Observation of the scarcity of interactions between the 49 genes that all were either up- or down-regulated in DN was unexpected. It is rational to assume that in the actual network between the genes related to DN, not all genes are regulated in the level of mRNA and hence not detected in the mRNA microarray experiment. The absence of these genes makes the interaction network incomplete. Therefore, the PPI network was enriched by the addition of maximum 2 interacting nodes for each gene. This resulted in expansion of the network from 49 nodes to 137 nodes. Indeed, the added 88 genes were predicted to be interacting with the 49 initial genes based on previous knowledge. The PPI network of these 137 genes was constructed with the same parameters applied for the initial network (Fig. 3A).


Nodes with high centrality in protein interaction networks are responsible for driving signaling pathways in diabetic nephropathy.

Abedi M, Gheisari Y - PeerJ (2015)

Enrichment of the PPI network and selection of central nodes for pathway enrichment analysis can determine pathways essentially related to DN.The 49-node network was extended with maximum two interactive genes for each node. The initial nodes selected from the microarray experiment are depicted with red color and enriched nodes with black (A). In this expanded network, 34 genes were selected as nodes with high centrality. Pathway enrichment analysis with these “central genes” disclosed 62 highly connected pathways related to DN. Pathways with adjusted p-value <0.05 are shown (B).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4636410&req=5

fig-3: Enrichment of the PPI network and selection of central nodes for pathway enrichment analysis can determine pathways essentially related to DN.The 49-node network was extended with maximum two interactive genes for each node. The initial nodes selected from the microarray experiment are depicted with red color and enriched nodes with black (A). In this expanded network, 34 genes were selected as nodes with high centrality. Pathway enrichment analysis with these “central genes” disclosed 62 highly connected pathways related to DN. Pathways with adjusted p-value <0.05 are shown (B).
Mentions: Observation of the scarcity of interactions between the 49 genes that all were either up- or down-regulated in DN was unexpected. It is rational to assume that in the actual network between the genes related to DN, not all genes are regulated in the level of mRNA and hence not detected in the mRNA microarray experiment. The absence of these genes makes the interaction network incomplete. Therefore, the PPI network was enriched by the addition of maximum 2 interacting nodes for each gene. This resulted in expansion of the network from 49 nodes to 137 nodes. Indeed, the added 88 genes were predicted to be interacting with the 49 initial genes based on previous knowledge. The PPI network of these 137 genes was constructed with the same parameters applied for the initial network (Fig. 3A).

Bottom Line: We found 49 genes to be variably expressed between the two groups.Interestingly, the central nodes were more informative for pathway enrichment analysis compared to all network nodes and also 49 differentially expressed genes.In conclusion, we here show that central nodes in protein interaction networks tend to be present in pathways that co-occur in a biological state.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Genetics and Molecular Biology, Isfahan University of Medical Sciences , Isfahan , Iran.

ABSTRACT
In spite of huge efforts, chronic diseases remain an unresolved problem in medicine. Systems biology could assist to develop more efficient therapies through providing quantitative holistic sights to these complex disorders. In this study, we have re-analyzed a microarray dataset to identify critical signaling pathways related to diabetic nephropathy. GSE1009 dataset was downloaded from Gene Expression Omnibus database and the gene expression profile of glomeruli from diabetic nephropathy patients and those from healthy individuals were compared. The protein-protein interaction network for differentially expressed genes was constructed and enriched. In addition, topology of the network was analyzed to identify the genes with high centrality parameters and then pathway enrichment analysis was performed. We found 49 genes to be variably expressed between the two groups. The network of these genes had few interactions so it was enriched and a network with 137 nodes was constructed. Based on different parameters, 34 nodes were considered to have high centrality in this network. Pathway enrichment analysis with these central genes identified 62 inter-connected signaling pathways related to diabetic nephropathy. Interestingly, the central nodes were more informative for pathway enrichment analysis compared to all network nodes and also 49 differentially expressed genes. In conclusion, we here show that central nodes in protein interaction networks tend to be present in pathways that co-occur in a biological state. Also, this study suggests a computational method for inferring underlying mechanisms of complex disorders from raw high-throughput data.

No MeSH data available.


Related in: MedlinePlus