Target Gene Screening And Evaluation Of Prognostic Values In Head And Neck Cancers By Bioinformatics Analysis

Background:In previous studies,the prognosis of head and neck tumors was limited to single-factor or single-gene studies.With the development of public databases represented by GEO,GCBI,HPA,and TCGA,high-throughput screening and head-neck selection was achieved through bioinformatics methods.It has become a new possibility to treat patients with individualized tumors based on genes associated with their prognosis.Objective:Screening target genes and evaluating prognostic values in head and neck cancers by integrated bioinformatics analysis Methods:1.The microarrays for head and neck cancers that meet the inclusion criteria were selected from the GEO and GCBI databases,and genes that were co-expressed in the head and neck malignancies on multiple sets of chips were analyzed.2.The gene data related to prognosis of head and neck cancer were obtained from the TCGA database,compared with the differentially expressed genesobtained from the GCBI database,and the two sets of common differentially expressed genes with good prognosis and poor prognosis were screened.3.To download the clinical data of the relevant patients from the HPA database and divide the patients into two groups including early and late stage(according to the TNM stage).To further analyze the relationship between the expression of genes indicating good/poor prognosis and tumor stage and survival probability,then select the genes with statistic difference.4.The expression data of the targeted genes in cancer tissues and adjacent normal tissues were downloaded from the TCGA database,the prognostic value of the single gene or multi-gene combinations were evaluated by the receiver operating characteristic curve(ROC curve).Results:1.Three microarrays(accession No.: GSE29330,GSE33205,and GSE9844)that met the inclusion criteria were selected from the GEO and GCBI databases.A total of 784 co-expressed genes were screened out,including 519 up-regulated genes and 265 down-regulated genes.2.Among the Co-expressed genes,we found 50 genes indicating poor prognosis and 27 genes indicating good prognosis.3.Eight genes were statistically associated with TNM staging and survival probability of SCCHN patients.The overexpression of ENDOU,NFIA,GIMAP6 or GIMAP7 indicated good prognosis;the high expression of USP14,P4HA1,PPFIA1 or PFDN2 indicated poor prognosis.4.Combination of ENDOU and NFIA or combination of ENDOU,NFIA,GIMAP6,and GIMAP7 had the best diagnostic performance,suggesting good prognosis;combination of USP14,P4HA1,PPFIA1 and PFDN2 have the mostdiagnostic efficacy,suggesting poor prognosis.Conclusion:1.ENDOU,NFIA,GIMAP6 and GIMAP7 were highly expressed in head and neck tumors,suggesting good prognosis;2.USP14,P4HA1,PPFIA1 and PFDN2 were highly expressed in head and neck tumors,indicating poor prognosis;3.Combination of ENDOU and NFIA or combination of ENDOU,NFIA,GIMAP6,and GIMAP7 had the best diagnostic performance,suggesting good prognosis;combination of USP14,P4HA1,PPFIA1 and PFDN2 have the most diagnostic efficacy,suggesting poor prognosis.Taken together,we have established a novel prognostic related gene profile of head and neck cancers by bioinformatics analysis.