Interactome-transcriptome analysis reveals the high centrality of genes differentially expressed in lung cancer tissues

Abstract

Motivation: Global protein interaction network (interactome) analysis provides an effective way to understand the relationships between genes. Through this approach, it was demonstrated that the essential genes in yeast tend to be highly connected as well as connected to other highly connected genes. This is in contrast to the genes that are not essential, which share neither of these properties. Using a similar interactome-transcriptome approach, the topological features in the interactome of differentially expressed genes in lung squamous cancer tissues are assessed.

Results: This analysis reveals that the genes that are differentially elevated, as obtained from the microarray gene profiling data, in cancer are well connected, whereas the suppressed genes and randomly selected ones are less so. These results support the notion that a topological analysis of cancer genes using protein interaction data will allow the placement of the list of genes, often of the disparate nature, into the global, systematic context of the cell. The result of this type of analysis may provide the rationale for therapeutic targets in cancer treatment.

Fig. 1

Correlation of connectivity (links) versus the fractions of select genes with exactly l links. The genes with exactly l links were chosen from the protein network, and then the fraction of selected genes from this subset was calculated. (a) Upregulated genes (n = 360) in SCC of lung have Pearson's r, which demonstrates high positive correlation (r = 0.82). (b) Downregulated genes (n = 270) in SCC have slightly less correlation (r = 0.75). (c) Microarray probesets that match the genes in the protein network (n = 2137) show no correlation to link number (r = 0.06). Thus, using the genes on the microarray does not contribute to bias in the number of links for genes differentially expressed in SCC. (FG values of 1 and 0 are not excluded for r values.)

Fig. 2

k-core analysis of differentially expressed genes in SCC of lung. k is the number of iterative decomposition of the graph (k-core subgraph), starting from the outermost edge. Higher k represents the central placement of the subset of genes in the original graph (k= 1). ER is the degree to which genes from a k-core is under- (ER < 1) or over-represented (ER > 1) relative to the original graph. a: ER of the upregulated genes (n = 360) in SCC; b: ER of the downregulated genes (n = 270) in SCC; c, c + σ, c − σ: mean ER of 1000 randomly selected genes from the entire network (n = 360), plus or minus the standard deviation (±σ); d, d + σ, d − σ: mean ER and ±σ for the 1000 randomly selected genes (n = 360) present in the microarray used in the experiment that matches the genes in the graph (n = 2137).