Multiple differential expression networks identify key genes in rectal cancer

Cancer Biomark. 2016;16(3):435-44. doi: 10.3233/CBM-160582.

Abstract

Background: Rectal cancer is an important contributor to cancer mortality.

Objective: The objective of this paper is to identify key genes across three phenotypes (fungating, polypoid and polypoid & small-ulcer) of rectal cancer based on multiple differential expression networks (DENs).

Methods: Differential interactions and non-differential interactions were evaluated according to Spearman correlation coefficient (SCC) algorithm, and were selected to construct DENs. Topological analysis was performed for exploring hub genes in largest components of DENs. Key genes were denoted as intersections between nodes of DENs and rectal cancer associated genes from Genecards. Finally, we utilized hub genes to classify phenotypes of rectal cancer on the basis of support vector machines (SVM) methodology.

Results: We obtained 19 hub genes and total 12 common key genes of three largest components of DENs, and EGFR was the common element. The SVM results revealed that hub genes could classify phenotypes, and validated feasibility of DEN methods.

Conclusions: We have successfully identified significant genes (such as EGFR and UBC) across fungating, polypoid and polypoid & small-ulcer phenotype of rectal cancer. They might be potential biomarkers for classification, detection and therapy of this cancer.

Keywords: Rectal cancer; differential expression network; differential interactions; genes; non-differential interactions.

MeSH terms

  • Antigens, Neoplasm / genetics
  • Biomarkers, Tumor / genetics*
  • Cyclin-Dependent Kinase Inhibitor p21 / genetics
  • ErbB Receptors / genetics
  • Gene Expression Profiling
  • Gene Expression Regulation, Neoplastic*
  • Humans
  • Protein Interaction Domains and Motifs / genetics*
  • Rectal Neoplasms / genetics*
  • Rectal Neoplasms / mortality
  • Signal Transduction / genetics
  • Support Vector Machine

Substances

  • Antigens, Neoplasm
  • Biomarkers, Tumor
  • CDKN1A protein, human
  • Cyclin-Dependent Kinase Inhibitor p21
  • UBC antigen
  • EGFR protein, human
  • ErbB Receptors