Identification of genes affecting in Non – small cell lung cancer using machine learning techniques and bioinformatics tools

Background & Objective: Lung cancer is the second most common cancer after breast cancer and the maintrigger cause of death in women and men globally. Non-small cell lung cancer (NSCLC) accounts for ۸۵%of lung cancers. Because early detection of cancer plays a vital role in treatment, this study sought to identifygenes that could potentially be effective in early Non – small cell lung cancer screening.Material & Methods: Firstly, three micro-array datasets (GSE۱۹۸۷, GSE۴۴۰۷۷, and GSE۷۴۷۰۶) related tonon-small cell lung cancer were downloaded from the Gene Expression Omnibus (GEO). After integratingand bath effect removal of these datasets, Lasso logistic regression was used to extract important genes.Processing of all data was performed using the R statistical programming language. Also, Gene SetEnrichment Analysis (GSEA) was performed by Metascape bioinformatics tool to identify KEGG pathwaysand Gene Ontology Enrichment.Results: Finally, the introduced model selected ۱۵ genes (ACVRL۱, ANKRD۱, C۱۱orf۸۰, CA۴, EIF۱B,FGF۲, GRK۵, KLHL۱۸, LILRA۱, MME, SDC۱, STX۱۱, TMOD۱, TTN, WIF۱). The accuracy level of themodel was ۱۰۰%. These genes are related to the Wnt signaling pathway, which plays a significant role inNSCLC. Until now, seven genes (۴۷%) have been reported in biological studies as genes effective in NSCLC.Conclusion: With the use of machine learning techniques and bioinformatics tools, this study has introducednew genes that can serve as the target of early diagnosis or treatment of NSCLC.

کلیدواژه ها:

Non – small cell lung cancer ، Gene expression ، Gene selection ، Machine learning ، Lasso logistic regression