Integrative Analysis of Gene Expression Through One-Class Logistic Regression Machine Learning Identifies Stemness Features in Multiple Myeloma

Abstract

Tumor progression includes the obtainment of progenitor and stem cell-like features and the gradual loss of a differentiated phenotype. Stemness was defined as the potential for differentiation and self-renewal from the cell of origin. Previous studies have confirmed the effective application of stemness in a number of malignancies. However, the mechanisms underlying the growth and maintenance of multiple myeloma (MM) stem cells remain unclear. We calculated the stemness index for samples of MM by utilizing a novel one-class logistic regression (OCLR) machine learning algorithm and found that mRNA expression-based stemness index (mRNAsi) was an independent prognostic factor of MM. Based on the same cutoff value, mRNAsi could stratify MM patients into low and high groups with different outcomes. We identified 127 stemness-related signatures using weighted gene co-expression network analysis (WGCNA) and differential expression analysis. Functional annotation and pathway enrichment analysis indicated that these genes were mainly involved in the cell cycle, cell differentiation, and DNA replication and repair. Using the molecular complex detection (MCODE) algorithm, we identified 34 pivotal signatures. Meanwhile, we conducted unsupervised clustering and classified the MM cohorts into three MM stemness (MMS) clusters with distinct prognoses. Samples in MMS-cluster3 possessed the highest stemness fractions and the worst prognosis. Additionally, we applied the ESTIMATE algorithm to infer differential immune infiltration among the three MMS clusters. The immune core and stromal score were significantly lower in MMS-cluster3 than in the other clusters, supporting the negative relation between stemness and anticancer immunity. Finally, we proposed a prognostic nomogram that allows for individualized assessment of the 3- and 5-year overall survival (OS) probabilities among patients with MM. Our study comprehensively assessed the MM stemness index based on large cohorts and built a 34-gene based classifier for predicting prognosis and potential strategies for stemness treatment.

Machine Learning and Deep Learning Applications in Multiple Myeloma Diagnosis, Prognosis, and Treatment Selection

“…However, the processes triggering the proliferation and survival of MM stem cells are not clear. The stemness index for MM cells was evaluated by employing a new one-class logistic regression (OCLR) ML system, and it was recognized that mRNAsi was an independent predictor for MM [57], able to differentiate MM subjects into groups with different OS. A total of 127 stemness-correlated signatures employing weighted gene co-expression network analysis (WGCNA) were recognised.…”

“…In addition, the employ of the ESTIMATE algorithm to evaluate a different immune activity among the three MMS groups allowed to confirm a negative correlation between stemness and anti-MM immunity. Finally, they suggested a predictive nomogram that allows individualized evaluation of the three-and five-year OS possibilities [57].…”