Machine learning-based identification of lower grade glioma stemness subtypes discriminates patient prognosis and drug response

Glioma stem cells (GSCs) remodel their tumor microenvironment to sustain a supportive niche. Identification and stratification of stemness related characteristics in patients with glioma might aid the diagnosis treatment disease. In this study, we calculated mRNA index bulk single-cell RNA-sequencing datasets using machine learning methods investigated correlation between clinicopathological characteristics. A stemness-associated score (GSScore) was constructed multivariate Cox regression analysis. We also generated GSC cell line derived from patient diagnosed used lines validate performance GSScore predicting chemotherapeutic responses. Differentially expressed genes (DEGs) GSCs high low GSScores were cluster lower-grade (LGG) samples into three subtypes. Differences characteristics, including survival, copy number variations, mutations, microenvironment, immune responses, among LGG subtypes identified. Using methods, further identified as subtype predictors validated CGGA datasets. current that correlated response. Through score, novel classification associated predictors, which facilitate development precision therapy.