Guangru Xu1, Minghui Zhang1, Hongxing Zhu1, Jinhua Xu2. 1. Department of Oncology, People's Hospital of Pudong District, Shanghai University of Medicine & Health Sciences, ShangHai 201299, China. 2. Department of Oncology, People's Hospital of Pudong District, Shanghai University of Medicine & Health Sciences, ShangHai 201299, China. Electronic address: wuwuuwgdg@aliyun.com.
Abstract
OBJECTIVE: To screen the gene signature for distinguishing patients with high risks from those with low-risks for colon cancer recurrence and predicting their prognosis. METHODS: Five microarray datasets of colon cancer samples were collected from Gene Expression Omnibus database and one was obtained from The Cancer Genome Atlas (TCGA). After preprocessing, data in GSE17537 were analyzed using the Linear Models for Microarray data (LIMMA) method to identify the differentially expressed genes (DEGs). The DEGs further underwent PPI network-based neighborhood scoring and support vector machine (SVM) analyses to screen the feature genes associated with recurrence and prognosis, which were then validated by four datasets GSE38832, GSE17538, GSE28814 and TCGA using SVM and Cox regression analyses. RESULTS: A total of 1207 genes were identified as DEGs between recurrence and no-recurrence samples, including 726 downregulated and 481 upregulated genes. Using SVM analysis and five gene expression profile data confirmation, a 15-gene signature (HES5, ZNF417, GLRA2, OR8D2, HOXA7, FABP6, MUSK, HTR6, GRIP2, KLRK1, VEGFA, AKAP12, RHEB, NCRNA00152 and PMEPA1) were identified as a predictor of recurrence risk and prognosis for colon cancer patients. CONCLUSION: Our identified 15-gene signature may be useful to classify colon cancer patients with different prognosis and some genes in this signature may represent new therapeutic targets.
OBJECTIVE: To screen the gene signature for distinguishing patients with high risks from those with low-risks for colon cancer recurrence and predicting their prognosis. METHODS: Five microarray datasets of colon cancer samples were collected from Gene Expression Omnibus database and one was obtained from The Cancer Genome Atlas (TCGA). After preprocessing, data in GSE17537 were analyzed using the Linear Models for Microarray data (LIMMA) method to identify the differentially expressed genes (DEGs). The DEGs further underwent PPI network-based neighborhood scoring and support vector machine (SVM) analyses to screen the feature genes associated with recurrence and prognosis, which were then validated by four datasets GSE38832, GSE17538, GSE28814 and TCGA using SVM and Cox regression analyses. RESULTS: A total of 1207 genes were identified as DEGs between recurrence and no-recurrence samples, including 726 downregulated and 481 upregulated genes. Using SVM analysis and five gene expression profile data confirmation, a 15-gene signature (HES5, ZNF417, GLRA2, OR8D2, HOXA7, FABP6, MUSK, HTR6, GRIP2, KLRK1, VEGFA, AKAP12, RHEB, NCRNA00152 and PMEPA1) were identified as a predictor of recurrence risk and prognosis for colon cancerpatients. CONCLUSION: Our identified 15-gene signature may be useful to classify colon cancerpatients with different prognosis and some genes in this signature may represent new therapeutic targets.
Authors: Sandra A S Johnson; Justin J Lin; Christopher J Walkey; Michael P Leathers; Cristian Coarfa; Deborah L Johnson Journal: Oncotarget Date: 2017-07-25
Authors: Johnny Yu; Albertas Navickas; Hosseinali Asgharian; Bruce Culbertson; Lisa Fish; Kristle Garcia; John Paolo Olegario; Maria Dermit; Martin Dodel; Benjamin Hänisch; Yikai Luo; Ethan M Weinberg; Rodrigo Dienstmann; Robert S Warren; Faraz K Mardakheh; Hani Goodarzi Journal: Cancer Discov Date: 2020-06-08 Impact factor: 38.272