The Prognostic and Diagnostic Value ofC۱orf۱۷۴ in Colorectal Cancer

Background: One of the most deadly forms of cancer is colorectalcancer (CRC), emphasizing the importance of discoveringnew biomarkers that can detect the disease at earlier stages.To accomplish this, scientists employed genome-wide RNAand microRNA sequencing, coupled with bioinformatics andmachine learning techniques, to identify genes with differentialexpression (DEGs). Subsequently, these DEGs were validatedin an additional cohort of CRC patients.Materials and Methods: In this study, a multi-step approachwas employed to identify potential biomarkers for colorectalcancer (CRC). First, RNA sequencing data from ۶۳۱ cases wereobtained from the Cancer Genome Atlas (TCGA), and differentialexpression genes (DEGs) were identified using the DESeq package in R. Next, prognostic biomarkers were identified byanalyzing survival curves using Kaplan-Meier analysis. Machinelearning algorithms such as Deep learning, Decision Tree,and Support Vector Machine were then used to identify predictivebiomarkers. Additional analyses included evaluating molecularpathways, protein-protein interactions (PPI), co-expressionof DEGs, and correlation between DEGs and clinical data.The diagnostic value of the identified markers was assessed usingthe combioROC package. Finally, a candidate top-scoringgene was validated using Real-time PCR in CRC patients.Results: Through survival analysis, we discovered five novelprognostic genes, namely KCNK۱۳, C۱orf۱۷۴, CLEC۱۸A,SRRM۵, and GPR۸۹A. Our study also revealed that three miRNAs,namely mir-۱۹b-۱, mir-۳۲۶, and mir-۳۳۰, were upregulatedin advanced stages of the disease. Combining the genesC۱orf۱۷۴, AKAP۴, DIRC۱, SKIL, and Scan۲۹A۴ yielded diagnosticmarkers with high sensitivity, specificity, and AUCvalues of ۰.۹۰, ۰.۹۴, and ۰.۹۲, respectively, as indicated bythe combineROC curve analysis. Finally, we validated theC۱orf۱۷۴ gene in CRC patients.Conclusion: Machine learning algorithms can effectively identifydysregulated genes/miRNAs involved in the pathogenesisof diseases, aiding in the early detection of patients. Furthermore,our data supports the prognostic and diagnostic value ofthe C۱orf۱۷۴ gene in colorectal cancer.

کلیدواژه ها:

Colorectal Cancer ، TCGA ، Machine Learning ، PrognosticBiomarker ، Diagnostic Biomarker