Weighted gene co-expression network analysis of colorectal cancer liver metastasis genome sequencing data and screening of anti-metastasis drugs

Distinct molecular profiles drive multifaceted characteristics of colorectal cancer metastatic seeds

Metastasis of primary tumors remains a challenge for early diagnosis and prevention. The cellular properties and molecular drivers of metastatically competent clones within primary tumors remain unclear. Here, we generated 10-16 single cell-derived lines from each of three colorectal cancer (CRC) tumors to identify and characterize metastatic seeds. We found that intrinsic factors conferred clones with distinct metastatic potential and cellular communication capabilities, determining organ-specific metastasis. Poorly differentiated or highly metastatic clones, rather than drug-resistant clones, exhibited poor clinical prognostic impact. Personalized genetic alterations, instead of mutation burden, determined the occurrence of metastatic potential during clonal evolution. Additionally, we developed a gene signature for capturing metastatic potential of primary CRC tumors and demonstrated a strategy for identifying metastatic drivers using isogenic clones with distinct metastatic potential in primary tumors. This study provides insight into the origin and mechanisms of metastasis and will help develop potential anti-metastatic therapeutic targets for CRC patients.

Prediction and validation of GUCA2B as the hub-gene in colorectal cancer based on co-expression network analysis: In-silico and in-vivo study

BACKGROUND

Several serious attempts to treat colorectal cancer have been made in recent decades. However, no effective treatment has yet been discovered due to the complexities of its etiology.

METHODS

we used Weighted Gene Co-expression Network Analysis (WGCNA) to identify key modules, hub-genes, and mRNA-miRNA regulatory networks associated with CRC. Next, enrichment analysis of modules has been performed using Cluepedia. Next, quantitative real-time PCR (RT-qPCR) was used to validate the expression of selected hub-genes in CRC tissues.

RESULTS

Based on the WGCNA results, the brown module had a significant positive correlation (r = 0.98, p-value=9e-07) with CRC. Using the survival and DEGs analyses, 22 genes were identified as hub-genes. Next, three candidate hub-genes were selected for RT-qPCR validation, and 22 pairs of cancerous and non-cancerous tissues were collected from CRC patients referred to the Gastroenterology and Liver Clinic. The RT-qPCR results revealed that the expression of GUCA2B was significantly reduced in CRC tissues, which is consistent with the results of differential expression analysis. Finally, top miRNAs correlated with GUCA2B were identified, and ROC analyses revealed that GUCA2B has a high diagnostic performance for CRC.

CONCLUSIONS

The current study discovered key modules and GUCA2B as a hub-gene associated with CRC, providing references to understand the pathogenesis and be considered a novel candidate to CRC target therapy.

Identification of Target PTEN-Based miR-425 and miR-576 as Potential Diagnostic and Immunotherapeutic Biomarkers of Colorectal Cancer With Liver Metastasis

A major complication of colorectal cancer (CRC), one of the most common and fatal types of cancers, is secondary liver metastasis. For patients with this fate, there are very few biomarkers available in clinical application, and the disease remains incurable. Recently, increasing studies demonstrated that tumorigenesis and development are closely related to immune escape, indicating that the roles of immune-related indicators might have been neglected in the past in colorectal cancer liver metastases (CRLM). Here, we unveil that elevated miR-425 and miR-576 promote CRLM through inhibiting PTEN-mediated cellular immune function. Specifically, miR-425 and miR-576 were identified for their significant upregulation in CRLM compared with the primary CRC tissues based on GSE81581 (n = 8) and GSE44121 (n = 18) datasets. Besides, we determined that the two microRNAs (miRNAs) coparticipated in restraining P53 and transforming growth factor beta (TGF-β) signaling pathways associated with tumor metastasis, and both shortened the overall survival of the patients with metastatic susceptibility. Notably, in situ hybridization on relatively large samples of paired CRC tissues (n = 157) not only substantiated that the expression of miR-425 and miR-576 was dramatically upregulated in CRLM but also revealed that they were closely related to tumor deterioration, especially liver metastases. Moreover, we further confirmed that the combination of miR-425 and miR-576 was an effective predictive model for liver metastases and poor clinical outcomes. Mechanically, downregulated PTEN (GSE81558, n = 6) was verified to be a shared target of miR-425 and miR-576 acting as metastasis-related oncogenes, on account of the presence of binding sites (+2928-+2934 and +4371-+4378, respectively) and the collaborative suppression of P53/TGF-β signaling in CRLM, which was further confirmed in CRC cells (HCT116 and SW480) based on systematic molecular biology experiments. Importantly, the target PTEN was strongly associated with microsatellite instability, tumor microenvironment, and immune cell infiltration. Thus, we speculate that miR-425 and miR-576 are novel biomarkers for CRLM prevention and immunotherapy and upstream inhibitors of the PTEN-P53/TGF-β function axis.

Hypermethylation and Downregulation of UTP6 Are Associated With Stemness Properties, Chemoradiotherapy Resistance, and Prognosis in Rectal Cancer: A Co-expression Network Analysis

Background

To identify the hub genes associated with chemoradiotherapy resistance in rectal cancer and explore the potential mechanism.

Methods

Weighted gene co-expression network analysis (WGCNA) was performed to identify the gene modules correlated with the chemoradiotherapy resistance of rectal cancer.

Results

The mRNA expression of 31 rectal cancer patients receiving preoperative chemoradiotherapy was described in our previous study. Through WGCNA, we demonstrated that the chemoradiotherapy resistance modules were enriched for translation, DNA replication, and the androgen receptor signaling pathway. Additionally, we identified and validated UTP6 as a new effective predictor for chemoradiotherapy sensitivity and a prognostic factor for the survival of colorectal cancer patients using our data and the GSE35452 dataset. Low UTP6 expression was correlated with significantly worse disease-free survival (DFS), overall survival (OS), and event- and relapse-free survival both in our data and the R2 Platform. Moreover, we verified the UTP6 expression in 125 locally advanced rectal cancer (LARC) patients samples by immunohistochemical analysis. The results demonstrated that low UTP6 expression was associated with worse DFS and OS by Kaplan-Meier and COX regression model analyses. Gene set enrichment and co-expression analyses showed that the mechanism of the UTP6-mediated chemoradiotherapy resistance may involve the regulation of FOXK2 expression by transcription factor pathways.

Conclusion

Low expression of the UTP6 was found to be associated with chemoradiotherapy resistance and the prognosis of colorectal cancer possibly via regulating FOXK2 expression by transcription factor pathways.

CTHRC1 promotes liver metastasis by reshaping infiltrated macrophages through physical interactions with TGF-β receptors in colorectal cancer

Metastasis is a major cause of cancer-related deaths. Tumor-intrinsic properties can determine whether tumor metastasis occurs or not. Here, by comparing the gene expression patterns in primary colorectal cancer (CRC) patients with or without metastasis, we found that Collagen Triple Helix Repeat Containing 1 (CTHRC1) in primary CRC served as a metastasis-associated gene. Animal experiments verified that CTHRC1 secreted by CRC cells promoted hepatic metastasis, which was closely correlated with macrophage infiltration. Depletion of macrophages by liposomal clodronate largely abolished the promoting effect of CTHRC1 on CRC liver metastasis. Furthermore, we demonstrated that CTHRC1 modulated macrophage polarization to M2 phenotypes through TGF-β signaling. A mechanistic study revealed that CTHRC1 bound directly to TGF-β receptor II and TGF-β receptor III, stabilized the TGF-β receptor complex, and activated TGF-β signaling. The combination treatment of CTHRC1 monoclonal antibody and anti-PD-1 blocking antibody effectively suppressed CRC hepatic metastasis. Taken together, our data demonstrated that CTHRC1 is an intrinsic marker of CRC metastasis and further revealed that CTHRC1 promoted CRC liver metastasis by reshaping infiltrated macrophages through TGF-β signaling, suggesting that CTHRC1 could be a potential biomarker for the early prediction of and a therapeutic target of CRC hepatic metastasis.

Identification of LncRNAs Associated With FOLFOX Chemoresistance in mCRC and Construction of a Predictive Model

Oxaliplatin, fluorouracil plus leucovorin (FOLFOX) regimen is the first-line chemotherapy of patients with metastatic colorectal cancer (mCRC). However, studies are limited regarding long non-coding RNAs (lncRNAs) associated with FOLFOX chemotherapy response and prognosis. This study aimed to identify lncRNAs associated with FOLFOX chemotherapy response and prognosis in mCRC patients and to construct a predictive model. We analyzed lncRNA expression in 11 mCRC patients treated with FOLFOX chemotherapy before surgery (four sensitive, seven resistant) by Gene Array Chip. The top eight lncRNAs (AC007193.8, CTD-2008N3.1, FLJ36777, RP11-509J21.4, RP3-508I15.20, LOC100130950, RP5-1042K10.13, and LINC00476) for chemotherapy response were identified according to weighted correlation network analysis (WGCNA). A competitive endogenous RNA (ceRNA) network was then constructed. The crucial functions of the eight lncRNAs enriched in chemotherapy resistance were mitogen-activated protein kinase (MAPK) and proteoglycans signaling pathway. Receiver operating characteristic (ROC) analysis demonstrated that the eight lncRNAs were potent predictors for chemotherapy resistance of mCRC patients. To further identify a signature model lncRNA chemotherapy response and prognosis, the validation set consisted of 196 CRC patients from our center was used to validate lncRNAs expression and prognosis by quantitative PCR (qPCR). The expression of the eight lncRNAs expression between CRC cancerous and adjacent non-cancerous tissues was also verified in the validation data set to determine the prognostic value. A generalized linear model was established to predict the probability of chemotherapy resistance and survival. Our findings showed that the eight-lncRNA signature may be a novel biomarker for the prediction of FOLFOX chemotherapy response and prognosis of mCRC patients.

Effects of Egr1 on pancreatic acinar intracellular trypsinogen activation and the associated ceRNA network

Acute pancreatitis (AP) is a common digestive disorder with high morbidity and mortality. The present study aimed to investigate the expression of early growth response protein 1 (Egr1), and the effect of competing endogenous (ce)RNA network on trypsinogen activation. Pancreatic acinar intracellular trypsinogen activation (PAITA) is an important event in the early stage of AP; however, the underlying mechanisms remain unclear. The present study used taurolithocholic acid 3-sulfate (TLC-S)-treated AR42J cells (pancreatic cell line) to establish a PAITA model. A gene microarray and bioinformatics analysis was performed to identify the potential key targets in PAITA. The results demonstrated that Egr1, an important transcription factor, was significantly overexpressed in PAITA. In Egr1 small interfering (si)RNA-transfected cells, Egr1 expression was decreased and trypsinogen activation was significantly decreased compared with negative control siRNA-transfected cells, indicating that in TLC-S-induced PAITA, overexpression of Egr1 enhanced trypsinogen activation. A ceRNA network [mRNA-microRNA (miRNA/miR)-long non-coding (lnc)RNA] generated using the PAITA model revealed that the effects of Egr1 on PAITA may be regulated by multiple ceRNA pairs, and the lncRNAs (including NONRATT022624 and NONRATT031002) and miRNAs [including Rattus norvegicus (rno)-miR-214-3p and rno-miR-764-5p] included in the ceRNA pairs may serve roles in PAITA by regulating the expression of Egr1. The results of the present study may provide novel targets for researching the underlying mechanisms of, and developing treatments for AP.

FBXW4 Acts as a Protector of FOLFOX-Based Chemotherapy in Metastatic Colorectal Cancer Identified by Co-Expression Network Analysis

Background

FOLFOX chemotherapy is one of the most commonly used treatments for colorectal cancer (CRC) patients. However, the efficacy and tolerance of FOLFOX therapy varies between patients. The purpose of this study was to explore hub genes associated with primary chemotherapy-resistance and to explore the possible mechanisms involved from non-European patients.

Method

A weighted gene co-expression network was constructed to identify gene modules associated with chemotherapy resistance in mCRC from China.

Results

A Gene Array Chip was used to detect mRNA expression in 11 mCRC patients receiving preoperative FOLFOX chemotherapy. The immune response was associated with chemotherapy-resistance in microarray data. Through the use of WGCNA, we demonstrated that the crucial functions enriched in chemotherapy-resistance modules were cell proliferation, MAPK signaling pathways, and PI3K signaling pathways. Additionally, we identified and validated FBXW4 as a new effective predictor for chemotherapy sensitivity and a prognostic factor for survival of CRC patients by using our own data and GSE69657. Furthermore, a meta-analysis of 15 Gene Expression Omnibus–sourced datasets showed that FBXW4 messenger RNA levels were significantly lower in CRC tissues than in normal colon tissues. An analysis of the data from the R2: Genomics Analysis and Visualization Platform showed that low FBXW4 expression was correlated with a significantly worse event- and relapse-free survival. Gene set enrichment analysis showed that the mechanism of FBXW4-mediated chemotherapy resistance may involve the DNA replication signal pathway and the cell cycle.

Conclusion

FBXW4 is associated with chemotherapy resistance and prognosis of CRC probably by regulating DNA replication signaling pathways and the cell cycle.

Weighted gene co-expression network analysis identified MYL9 and CNN1 are associated with recurrence in colorectal cancer

Colorectal cancer (CRC) is one of the most common carcinomas and the fourth leading cause of cancer-related death worldwide. One of the obstacles in the successful treatment of CRC is a high rate of recurrence. We aimed to construct weighted gene co-expression network analysis (WGCNA) to identify key modules and hub genes in association with recurrence in CRC patients. We firstly used the microarray data, GSE41258, to construct a co-expression network and identify gene modules. Furthermore, protein and protein interaction (PPI) network was also performed to screen hub genes. To validate the hub genes, an independent dataset GSE17536 was used for survival analyses. Additionally, another two databases were also performed to investigate the survival rates and expression levels of hub genes. Gene set enrichment analyses (GSEA) combined with gene ontology (GO) were performed to further explore function and mechanisms. In our study, the midnightblue module was identified to be significant, 15 hub genes were screened, four of which were identified as hub nodes in the PPI network. In the test dataset, we found higher expression of MYL9 and CNN1 were significantly associated with shorter survival time of CRC patients. GO analyses showed that MYL9 and CNN1 were enriched in “muscle system process” and “cytoskeletal protein binding”. GSEA found the two hub genes were enriched in “pathways in cancer” and “calcium signaling pathway”. In conclusion, our study demonstrated that MYL9 and CNN1 were hub genes associated with the recurrence of CRC, which may contribute to the improvement of recurrence-free survival time of CRC patients.

Optimal Bayesian Filtering for Biomarker Discovery: Performance and Robustness

Optimal Bayesian feature filtering (OBF) is a fast and memory-efficient algorithm that optimally identifies markers with distributional differences between treatment groups under Gaussian models. Here, we study the performance and robustness of OBF for biomarker discovery. Our contributions are twofold: (1) we examine how OBF performs on data that violates modeling assumptions, and (2) we provide guidelines on how to set input parameters for robust performance. Contribution (1) addresses an important, relevant, and commonplace problem in computational biology, where it is often impossible to validate an algorithm's core assumptions. To accomplish both tasks, we present a battery of simulations that implement OBF with different inputs and challenge each assumption made by OBF. In particular, we examine the robustness of OBF with respect to incorrect input parameters, false independence, imbalanced sample size, and we address the Gaussianity assumption by considering performance on an extensive family of non-Gaussian distributions. We address advantages and disadvantages between different priors and optimization criteria throughout. Finally, we evaluate the utility of OBF in biomarker discovery using acute myeloid leukemia (AML) and colon cancer microarray datasets, and show that OBF is successful at identifying well-known biomarkers for these diseases that rank low under moderated t-test.

Discovery of core genes in colorectal cancer by weighted gene co-expression network analysis

The aim of the present study was to investigate the interactions among messenger RNAs (mRNAs), microRNAs (miRNAs), and long noncoding RNAs (lncRNAs) in colorectal cancer (CRC), in order to examine its underlying mechanisms. The raw gene expression data was downloaded from the Gene Expression Omnibus (GEO) database. An online tool, GEO2R, which is based on the limma package, was used to identify differentially expressed genes. The co-expression between lncRNAs and mRNAs was identified utilizing the weighted gene co-expression analysis package of R to construct a coding non-coding (CNC) network. The function of the genes in the CNC network was determined by performing Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathways enrichment analysis. The interactions among miRNAs, mRNAs and lncRNAs were predicted using Lncbase and mirWalk to construct the competing endogenous RNA (ceRNA) network. The expression of the genes involved in the ceRNA network was further validated in The Cancer Genome Atlas dataset. A total of 3,183 dysregulated mRNAs, 78 dysregulated miRNAs and 2,248 dysregulated lncRNAs were screened in two GEO datasets. Combined with the results of the dysregulated genes, 169 genes were selected to construct the CNC network. 'p53 signaling pathway' and the 'cell cycle' were the most significant enriched pathways in the genes involved in the CNC network. Finally, a validated ceRNA network composed of 2 lncRNAs (MIR22HG and RP11-61I13.3), 5 miRNAs (hsa-miR-765, hsa-miR-198, hsa-miR-125a-3p, hsa-miR-149-3p and hsa-miR-650) and 5 mRNAs (ANK2, BTK, GBP2, PCSK5 and PDK4) was obtained. In conclusion, MIR22HG may regulate PCSK5, BTK and PDK4, and RP11-61I13.3 may regulate the ANK2, GBP2, PCSK5 through sponging miRNAs to act on the progression of CRC, and the potential function of these genes have been revealed. However, the diagnostic and prognostic value of these genes requires further validation.

Identification of specific modules and significant genes associated with colon cancer by weighted gene co‑expression network analysis

Colon cancer is one of the most commonly diagnosed malignancies and is a leading cause of cancer‑associated mortality. The aim of the present study was to investigate the molecular mechanisms underlying colon cancer and identify potentially significant genes associated with the disease using weighted gene co‑expression network analysis (WGCNA). The test datasets used were downloaded from The Cancer Genome Atlas (TCGA) database. WGCNA was applied to analyze microarray data obtained from colon adenocarcinoma samples to identify significant modules and highly associated genes. A gene co‑expression network was constructed and different gene modules were selected. Functional and pathway enrichment analyses were performed to investigate the molecular mechanisms of colon cancer. In addition, highly connected hub genes associated with the most significant module were selected for further analysis. Nine specific modules associated with colon cancer were identified, of which the turquoise module was observed to exhibit the greatest association with the disease. Pathway enrichment analysis of the turquoise module suggested that genes in the turquoise module were associated with 'RNA polymerase' and 'purine metabolism'. Furthermore, gene ontology enrichment analysis revealed the top 30 hub genes with a higher degree in the turquoise module, such as σ‑non‑opioid intracellular receptor 1, transmembrane protein 147  TMEM147) and carbamoyl‑phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase, were predominantly enriched in the biological processes 'translation' and 'gene expression'. Experimental verification demonstrated that the expression of TMEM147 in colon cancer was significantly increased compared with the control. Therefore, the results suggested that genes associated with RNA polymerase and the purine metabolic pathways may be substantially involved in the pathogenesis of colon cancer. Furthermore, TMEM147 may represent a biomarker for colon cancer.

MicroRNAs mediate liver transcriptome changes upon soy diet intervention in mice

Soy‐based diets have triggered the interest of the research community due to their beneficial effects on a wide variety of pathologies like breast and prostate cancer, diabetes and atherosclerosis. However, the molecular details underlying these effects are far from being completely understood; several recent attempts have been made to elucidate the soy‐induced liver transcriptome changes in different animal models. Here we used Next Generation Sequencing to identify a set of microRNAs specifically modulated by short‐term soy‐enriched diet in young male mice and estimated their impact on the liver transcriptome assessed by microarray. Clustering and topological community detection (CTCD) network analysis of STRING generated interactions of transcriptome data led to the identification of four topological communities of genes characteristically altered and putatively targeted by microRNAs upon soy diet intervention.

Identification of gene expression profiles and key genes in subchondral bone of osteoarthritis using weighted gene coexpression network analysis

Osteoarthritis (OA) significantly influences the quality life of people around the world. It is urgent to find an effective way to understand the genetic etiology of OA. We used weighted gene coexpression network analysis (WGCNA) to explore the key genes involved in the subchondral bone pathological process of OA. Fifty gene expression profiles of GSE51588 were downloaded from the Gene Expression Omnibus database. The OA-associated genes and gene ontologies were acquired from JuniorDoc. Weighted gene coexpression network analysis was used to find disease-related networks based on 21756 gene expression correlation coefficients, hub-genes with the highest connectivity in each module were selected, and the correlation between module eigengene and clinical traits was calculated. The genes in the traits-related gene coexpression modules were subject to functional annotation and pathway enrichment analysis using ClusterProfiler. A total of 73 gene modules were identified, of which, 12 modules were found with high connectivity with clinical traits. Five modules were found with enriched OA-associated genes. Moreover, 310 OA-associated genes were found, and 34 of them were among hub-genes in each module. Consequently, enrichment results indicated some key metabolic pathways, such as extracellular matrix (ECM)-receptor interaction (hsa04512), focal adhesion (hsa04510), the phosphatidylinositol 3'-kinase (PI3K)-Akt signaling pathway (PI3K-AKT) (hsa04151), transforming growth factor beta pathway, and Wnt pathway. We intended to identify some core genes, collagen (COL)6A3, COL6A1, ITGA11, BAMBI, and HCK, which could influence downstream signaling pathways once they were activated. In this study, we identified important genes within key coexpression modules, which associate with a pathological process of subchondral bone in OA. Functional analysis results could provide important information to understand the mechanism of OA.

MicroRNA-204 Inhibits the Growth and Motility of Colorectal Cancer Cells by Downregulation of CXCL8

Among all of the miRNAs, miR-204 has gained considerable attention in the field of cancer research. This study aimed to reveal the detailed functions and the underlying mechanism of miR-204 in colorectal cancer (CRC) cells. The expressions of miR-204 in CRC tumor tissues and cell lines were monitored. Expressions of miR-204 and CXCL8 in Caco-2 and HT-29 cells were altered by transfection, and then cell viability, apoptosis, migration, invasion, EMT-related protein expression, and PI3K/AKT/mTOR pathway protein expression were assessed. We found that miR-204 was expressed at low levels in CRC tumor tissues and cell lines when compared to their normal controls. miR-204 overexpression reduced the viability, migration, and invasion of Caco-2 and HT-29 cells while significantly inducing apoptosis. miR-204 overexpression upregulated E-cadherin expression and downregulated N-cadherin and vimentin expressions. CXCL8 was a target of miR-204, and miR-204 suppression could not increase cell viability, migration, invasion, and EMT procedure when CXCL8 was silenced. Moreover, miR-204 overexpression decreased the phosphorylated levels of PI3K, AKT, and mTOR. The increased phosphorylations of PI3K, AKT, and mTOR, and the upregulation of CXCL8 induced by miR-204 suppression were all abolished by the addition of LY294002 and AZD8055 (inhibitors of PI3K/AKT and mTOR, respectively). To conclude, we demonstrated a tumor-suppressive miRNA in CRC cell lines, miR-204, which is poorly expressed in CRC tissues and cell lines. miR-204 exerted antigrowth, antimigration, anti-invasion, and anti-EMT activities, which might be via deactivating the PI3K/AKT/mTOR pathway and repressing CXCL8 expression.

The integration of weighted gene association networks based on information entropy

Constructing genome scale weighted gene association networks (WGAN) from multiple data sources is one of research hot spots in systems biology. In this paper, we employ information entropy to describe the uncertain degree of gene-gene links and propose a strategy for data integration of weighted networks. We use this method to integrate four existing human weighted gene association networks and construct a much larger WGAN, which includes richer biology information while still keeps high functional relevance between linked gene pairs. The new WGAN shows satisfactory performance in disease gene prediction, which suggests the reliability of our integration strategy. Compared with existing integration methods, our method takes the advantage of the inherent characteristics of the component networks and pays less attention to the biology background of the data. It can make full use of existing biological networks with low computational effort.

A five-gene based risk score with high prognostic value in colorectal cancer

Colorectal cancer (CRC) is one of the most frequently occurring malignancies worldwide. The outcomes of patients with similar clinical symptoms or at similar pathological stages remain unpredictable. This inherent clinical diversity is most likely due to the genetic heterogeneity. The present study aimed to create a predicting tool to evaluate patient survival based on genetic profile. Firstly, three Gene Expression Omnibus (GEO) datasets (GSE9348, GSE44076 and GSE44861) were utilized to identify and validate differentially expressed genes (DEGs) in CRC. The GSE14333 dataset containing survival information was then introduced in order to screen and verify prognosis-associated genes. Of the 66 DEGs, the present study screened out 46 biomarkers closely associated to patient overall survival. By Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathway analysis, it was demonstrated that these genes participated in multiple biological processes which were highly associated with cancer proliferation, drug-resistance and metastasis, thus further affecting patient survival. The five most important genes, MET proto-oncogene, receptor tyrosine kinase, carboxypeptidase M, serine hydroxymethyltransferase 2, guanylate cyclase activator 2B and sodium voltage-gated channel a subunit 9 were selected by a random survival forests algorithm, and were further made up to a linear risk score formula by multivariable cox regression. Finally, the present study tested and verified this risk score within three independent GEO datasets (GSE14333, GSE17536 and GSE29621), and observed that patients with a high risk score had a lower overall survival (P<0.05). Furthermore, this risk score was the most significant compared with other predicting factors including age and American Joint Committee on Cancer stage, in the model, and was able to predict patient survival independently and directly. The findings suggest that this survival associated DEGs-based risk score is a powerful and accurate prognostic tool and is promisingly implemented in a clinical setting.

Identification of two novel biomarkers of rectal carcinoma progression and prognosis via co-expression network analysis

mRNA expression profiles provide important insights on a diversity of biological processes involved in rectal carcinoma (RC). Our aim was to comprehensively map complex interactions between the mRNA expression patterns and the clinical traits of RC. We employed the integrated analysis of five microarray datasets and The Cancer Genome Atlas rectal adenocarcinoma database to identify 2118 consensual differentially expressed genes (DEGs) in RC and adjacent normal tissue samples, and then applied weighted gene co-expression network analysis to parse DEGs and eight clinical traits in 66 eligible RC samples. A total of 16 co-expressed gene modules were identified. The green-yellow and salmon modules were most appropriate to the pathological stage (R = 0.36) and the overall survival (HR =13.534, P = 0.014), respectively. A diagnostic model of the five pathological stage hub genes (SCG3, SYP, CDK5R2, AP3B2, and RUNDC3A) provided a powerful classification accuracy between localized RC and non-localized RC. We also found increased Secretogranin III (SCG3) expression with higher pathological stage and poorer prognosis in the test and validation set. The increased Homer scaffolding protein 2 (HOMER2) expression with the favorable survival prediction efficiency significantly correlated with the markedly reduced overall survival of RC patients and the higher pathological stage during the test and validation set. Our findings indicate that the SCG3 and HOMER2 mRNA levels should be further evaluated as predictors of pathological stage and survival in patients with RC.