Predictive Biomarkers and Molecular Subtypes in DLBCL: Insights from PCD Gene Expression and Machine Learning

The data consists of gene expression profiles from five DLBCL patient cohorts (GSE132929, GSE43677, GSE156309, GSE190847, and GSE12453) and 18 programmed cell death (PCD)-related genes. The data include results from consensus clustering that divided the patients into two subgroups (C1 and C2), with C2 representing a higher-risk group. The dataset also includes immune infiltration data, gene set variation analysis (GSVA) of relevant pathways, and predictive models built using 12 different machine learning algorithms. Validation through transcriptome sequencing of DLBCL cell lines and normal B-lymphocyte cell lines was also conducted.