Oncogenes and tumor suppressor genes: comparative genomics and network perspectives

Abstract

Background: Defective tumor suppressor genes (TSGs) and hyperactive oncogenes (OCGs) heavily contribute to cell proliferation and apoptosis during cancer development through genetic variations such as somatic mutations and deletions. Moreover, they usually do not perform their cellular functions individually but rather execute jointly. Therefore, a comprehensive comparison of their mutation patterns and network properties may provide a deeper understanding of their roles in the cancer development and provide some clues for identification of novel targets.

Results: In this study, we performed a comprehensive survey of TSGs and OCGs from the perspectives of somatic mutations and network properties. For comparative purposes, we choose five gene sets: TSGs, OCGs, cancer drug target genes, essential genes, and other genes. Based on the data from Pan-Cancer project, we found that TSGs had the highest mutation frequency in most tumor types and the OCGs second. The essential genes had the lowest mutation frequency in all tumor types. For the network properties in the human protein-protein interaction (PPI) network, we found that, relative to target proteins, essential proteins, and other proteins, the TSG proteins and OCG proteins both tended to have higher degrees, higher betweenness, lower clustering coefficients, and shorter shortest-path distances. Moreover, the TSG proteins and OCG proteins tended to have direct interactions with cancer drug target proteins. To further explore their relationship, we generated a TSG-OCG network and found that TSGs and OCGs connected strongly with each other. The integration of the mutation frequency with the TSG-OCG network offered a network view of TSGs, OCGs, and their interactions, which may provide new insights into how the TSGs and OCGs jointly contribute to the cancer development.

Conclusions: Our study first discovered that the OCGs and TSGs had different mutation patterns, but had similar and stronger protein-protein characteristics relative to the essential proteins or control proteins in the whole human interactome. We also found that the TSGs and OCGs had the most direct interactions with cancer drug targets. The results will be helpful for cancer drug target identification, and ultimately, understanding the etiology of cancer and treatment at the network level.

Figure 1

Percentage comparison of Pan-Cancer samples mutated in five gene sets (A and B) and across 12 tumor types (C and D). In Figure D, one star indicates a P-value less than 0.05 based on the Kolmogorov-Smirnov (K-S) test between the two gene sets. The star color indicates the corresponding gene sets. For example, in BLCA, the top of the TSG bar has four stars, which indicates that the percentage of samples with mutations in the TSG gene set was significantly higher than that of OCG genes (blue star), target genes (red star), essential genes (green star), and other genes (gray star). 'TSG' represents the tumor suppressor genes, 'OCG' represents the oncogenes, 'Target' represents the genes encoding cancer drug targets, 'Essential' represents the essential genes, and 'Other' represents the other genes with mutation data that are part of the PPI data.

Figure 2

Comparison of degree and betweenness of five protein sets. A) Degree distribution. B) Summary of the average degree and the corresponding P-values of the Kolmogorov-Smirnov (K-S) tests for any two protein sets. C) Betweenness distribution. D) Summary of the average betweenness (1.0 × 10³) and the corresponding P-values of the K-S tests for any two protein sets.

Figure 3

Distribution of clustering coefficient of five protein sets. The inserted table summarizes the average value of clustering coefficient for each protein set and the corresponding P-values based on the Kolmogorov-Smirnov (K-S) tests for any two protein sets.

Figure 4

Distribution of shortest-path distance from five protein sets to the other nodes in human protein-protein interaction network. The inserted table summarizes the average value of shortest-path distance from each protein set to the rest nodes in human protein-protein interaction network and the corresponding P-values based on the Kolmogorov-Smirnov (K-S) tests for any two protein sets.

Figure 5

Network-based relationship between target proteins to other four protein sets. A) Distribution of shortest-path distance of five gene sets. B) Protein proportion at the shortest-path distance 1 and 2 from target proteins to TSG proteins or OCG proteins.

Figure 6

TSG-OCG network. Node color indicates the different protein sets: red for TSG proteins, blue for OCG proteins, and green for linkers that could link TSG proteins and OCG proteins. Edge color indicates protein-protein interaction among different protein sets: red for the interactions among TSG proteins, blue for the interactions among OCG proteins, dark green for the interactions between TSG proteins and OCG proteins, and gray for the interactions between linkers and OCG proteins or TSG proteins. Node size is corresponds to the mutation frequency in Pan-Cancer samples. The larger the node, the higher the frequency was.