Clinical information of CRC patients from the TCGA
Database was available in 539 cases. The clinical characteristics including age, gender, tumor stage ,and lymph node metastasis of CRC patients are presented in Table 1. The median age of the cohort was 68 years(range: 31–90 years).
Table 1
The clinicopathological characteristics of CRC patients.
Age | <=65 | 223 | 41.37% |
> 65 | 316 | 58.63% |
gender | Female | 256 | 47.50% |
Male | 283 | 52.50% |
Tumor Stage | I | 91 | 16.88% |
II | 197 | 36.55% |
III | 156 | 28.94% |
IV | 76 | 14.10% |
Lymph node | N0 | 305 | 56.59% |
N1 | 125 | 23.19% |
N2 | 105 | 19.48% |
M | M0 | 394 | 73.10% |
M1 | 76 | 14.10% |
N = 539 |
Missing values wasn’t presented in Table. |
Identification of differentially expressed genes
A total of 514 mRNA samples were enrolled in our study, including 473 colon carcinoma tissues and 41 matched normal tissues. To pick out the differentially expressed genes with normal issues and CRC, we conducted differentially expression analysis. We filtered low expression genes. To date,719 mRNAs were found to be significantly upregulated and downregulated in CRC. For differentially expressed genes, we identified 375 upregulated genes and 344 downregulated genes. The heatmap showed 50 samples by the differentially expressed genes (Fig. 1). The volcano plot showed the relationship between FDR(false discovery rate) and Fold Change(FC) in all differentially expressed genes(Fig. 2).It showed normal and cancer group could be clearly distinguished by the differentially expressed genes.
Functional enrichment analysis of differentially expressed genes
To further investigate the biological functions of these differentially expressed genes, we conducted functional enrichment analysis. The BP(biological process )terms, CC(cellular components)terms and MF(molecular functions) terms were exhibited in Fig. 3A and Table 2. Proteolysis,protein heterodimerization activity and extracellular exosome were the main enriched GO(gene ontology) terms. KEGG(Kyoto Encyclopedia of Genes and Genomes) pathways were displayed in Fig. 3B. The main enriched pathways was cytokine-cytokine receptor interaction, alcoholism and systemic lupus erythematosus.
Protein-protein interaction (PPI) network and identification of hub genes
The PPI network of differentially expressed genes contained 715 nodes and 1186 edges. We displayed the sub-network with most nodes and edges in Fig. 4. The average node degree is 3.32. Expected number of edges is 532.PPI enrichment p-value < 1.0− 16. According to the PPI network, my network has significantly more interactions than expected what means functional enrichments in my network. The degree of these genes was shown in Table 3. Top 10 genes (degree >/=25) were selected as hub genes (Fig. 5),including:GNG4,PPBP,NMU,LPAR1,CXCL1,HIST1H2BJ,CXCL8,PYY,HIST1H2AD and SST.GNG4 could be the key genes in colorectal cancer using bioinformatics analysis.
GNG4 expression difference validation based on the Oncomine database
Four CRC datasets in the Oncomine database were adopted for the validation of gng4 over-expression(Skrzypczak Colorectal2 Statistics[http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE20916];Kaiser Colon Statistics[http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE 5206];Hong Colorectal Statistics [http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE9348];Skrzypczak Colorectal Statistics[http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE 20916]),which was in accordance with the results of the meta-analysis (Fig. 6A,6B,6C,6D). The P-values for these datasets were 1.13− 10, 8.31− 7, 7.09− 9 and3.53− 5,respectively. However, the remaining datasets did not show a statistically significant difference in GNG4 expression. The results of GNG4 copy number in TCGA Colorectal 2 Grouped by overall survival status was showed in Fig. 7 (http://tcga-data.nci.nih.gov/tcga/). The median rank is 1219.5 after comparison of GNG4 across 4 analyses(P = 4.19-7,Fig. 11) by meta-analysis. Shows that GNG4 overexpression was statistically significant. GNG4 co-expression results were shown in Fig. 8.
Analysis of clinical magnitude
The clinical magnitude of GNG4 in CRC was assessed on the basis of TCGA clinical data. A total of 539 patient samples with CRC were retrieved from a cohort of TCGA database. The association of gene expression with tumor stage, lymph node, metastasis, age, gender was analyzed. It was significantly different between with stage I and IV, II and III, II and IV,III and IV (P = 0.0015,0.029,3.9− 05,0.00074,0.01,respectively) (Fig. 9A). It was a relationship with stage. It was a relationship with lymph node, where the frequency of GNG4 was significantly different in patients with N1,N2 than N0(P = 0.0025, 0.037)( Fig. 9B). We observed that GNG4 was highly expressed in metastasis patients(P = 9.5− 05, Fig. 9C).Taken together, GNG4 was associated with stage, metastasis and lymph node. In other words, GNG4 maybe a prognostic molecule. GNG4 expression did not show a statistically significant difference in in age and gender(P = 0.16,0.1, respectively) (Fig. 9D,9E). There wasn’t statistically different between I and II,I and III may due to the small number of stage I patients. There wasn’t statistically different in survival curve of GNG4 expression between low level group and high group(P = 0.96641, Fig. 10).We should collect clinical patients to validate.