Construction of a non-negative matrix factorization model of immunogenic cell death-related genes in lung adenocarcinoma and analysis of survival prognosis

Abstract

PurposeTo explore the effectiveness of the model based on non-negative matrix factorization (NMF), analyze the tumor microenvironment and immune microenvironment for evaluating the prognosis of lung adenocarcinoma, establish a risk model, and screen independent prognostic factors. MethodsDownloading the transcription data files and clinical information files of lung adenocarcinoma from TCGA database and GO database, the R software was used to establish the NMF cluster model, and then the survival analysis between groups, tumor microenvironment analysis, and immune microenvironment analysis was performed according to the NMF cluster result. R software was used to construct prognostic models and calculate risk scores. Survival analysis was used to compare survival differences between different risk score groups. ResultsTwo ICD subgroups were established according to the NMF model. The survival of the ICD low-expression subgroup was better than that of the ICD high-expression subgroup. Univariate COX analysis screened out HSP90AA1, IL1, and NT5E as prognostic genes, and the prognostic model established on this basis has clinical guiding significance. ConclusionThe model based on NMF has the prognostic ability for lung adenocarcinoma, and the prognostic model of ICD-related genes has a certain guiding significance for survival.