Glioma is a malignancy with the highest mortality in central nervous system disorders. Here, we implemented the computational tools based on CRISPR/Cas9 to predict the clinical outcomes and biological characteristics of low-grade glioma (LGG). The transcriptional expression profiles and clinical phenotypes of LGG patients were retrieved from The Cancer Genome Atlas and Chinese Glioma Genome Atlas. The CERES algorithm was used to screen for LGG-lethal genes. Cox regression and random survival forest were adopted for survival-related gene selection. Nonnegative matrix factorization distinguished patients into different clusters. Single-sample gene set enrichment analysis was employed to create a novel CRISPR/Cas9 screening potential index (CCSPI), and patients were stratified into low- and high-CCSPI groups. Survival analysis, area under the curve values (AUCs), nomogram, and tumor microenvironment exploration were included for the model validation. A total of 20 essential genes in LGG were used to classify patients into two clusters and construct the CCSPI system. High-CCSPI patients were associated with a worse prognosis of both training and validation set (p < 0.0001) and higher immune fractions than low-CCSPI individuals. The CCSPI system had a promising performance with 1-, 3-, and 5-year AUCs of 0.816, 0.779, 0.724, respectively, and the C-index of the nomogram model reached 0.743 (95% CI = 0.725–0.760). Immune-infiltrating cells and immune checkpoints such as PD-1/PD-L1 and POLD3 were positively associated with CCSPI. In conclusion, the CCSPI had prognostic value in LGG, and the model will deepen our cognition of the interaction between the CNS and immune system in different LGG subtypes.

Background: An imbalance of redox homeostasis participates in tumorigenesis, proliferation, and metastasis, which results from the production of reactive oxygen species (ROS). However, the biological mechanism and prognostic significance of redox-associated messenger RNAs (ramRNAs) in lung adenocarcinoma (LUAD) still remain unclear.Methods: Transcriptional profiles and clinicopathological information were retrieved from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) of LUAD patients. A total of 31 overlapped ramRNAs were determined, and patients were separated into three subtypes by unsupervised consensus clustering. Biological functions and tumor immune-infiltrating levels were analyzed, and then, differentially expressed genes (DEGs) were identified. The TCGA cohort was divided into a training set and an internal validation set at a ratio of 6:4. Least absolute shrinkage and selection operator regression were used to compute the risk score and determine the risk cutoff in the training set. Both TCGA and GEO cohort were distinguished into a high-risk or low-risk group at the median cutoff, and then, relationships of mutation characteristics, tumor stemness, immune differences, and drug sensitivity were investigated.Results: Five optimal signatures (ANLN, HLA-DQA1, RHOV, TLR2, and TYMS) were selected. Patients in the high-risk group had poorer prognosis, higher tumor mutational burden, overexpression of PD-L1, and lower immune dysfunction and exclusion score compared with the low-risk group. Cisplatin, docetaxel, and gemcitabine had significantly lower IC50 in the high-risk group.Conclusion: This study constructed a novel predictive signature of LUAD based on redox-associated genes. Risk score based on ramRNAs served as a promising biomarker for prognosis, TME, and anti-cancer therapies of LUAD.