The landscape of gene co-expression modules correlating with prognostic genetic abnormalities in AML

Abstract Background The heterogenous cytogenetic and molecular variations were harbored by AML patients, some of which are related with pathogenesis clinical outcomes. We aimed to uncover the intrinsic expression profiles correlating prognostic genetic abnormalities WGCNA. Methods downloaded dataset from BeatAML, TCGA GEO database. Using R (version 4.0.2) ‘WGCNA’ package, co-expression modules ELN2017 markers identified (R 2 ? 0.4, p < 0.01). ORA detected enriched pathways for key modules. patients in cohort randomly assigned into training set (50%) testing (50%). LASSO penalized regression analysis was employed build prediction model, fitting OS level hub genes ‘glmnet’ package. Then independent validation sets (GSE12417 GSE37642) used validate diagnostic utility accuracy model. Results A total 37 gene 973 BeatAML cohort. found that 3 significantly correlated (the ‘lightyellow’ module NPM1 mutation, ‘saddlebrown’ RUNX1 ‘lightgreen’ TP53 mutation). revealed mainly DNA-binding transcription factor activity activation HOX genes. immune response process. And predominantly mitosis cell cycle LASSO- 6 (NFKB2, NEK9, HOXA7, APRC5L, FAM30A LOC105371592) non-zero coefficients. risk score generated 6-gene associated stratification, relapsed disease, prior MDS history. 5-year AUC model 0.822 0.824 sets, respectively. Moreover, robust when it (5-year 0.743–0.79). Conclusions established network signature recommended AML. survival developed validated multiple datasets.