Integrative analysis of multi-omics data reveals a pseudouridine-related lncRNA signature for prediction of glioma prognosis and chemoradiotherapy sensitivity
Yanbo Yang,Fei Wang,Haiying Teng,Chuanpeng Zhang,Yulian Zhang,Pengyu Chen,Quan Li,Xiuji Kan,Zhouqing Chen,Zhong Wang,Yanbing Yu
DOI: https://doi.org/10.1016/j.compbiomed.2023.107428
IF: 7.7
2023-11-01
Computers in Biology and Medicine
Abstract:BACKGROUND: Glioblastoma is the most common type of glioma with a high incidence and poor prognosis, and effective medical treatment remains challenging. Pseudouridine (Ψ) is the first post-transcriptional modification discovered and one of the most abundant modifications to RNA. However, the prognostic value of Ψ-related lncRNAs (ΨrLs) for glioma patients has never been systematically evaluated. This study aims to construct a risk model based on ΨrLs signature and to validate the predictive efficiency of the model.METHOD: Transcriptomic data, genomic data, and relevant clinical data of glioma patients were extracted from the Cancer Genome Atlas (TCGA) and the Chinese Glioma Genome Atlas (CGGA). ΨrLs with significant correlation with Ψ-related genes were identified, and univariate Cox regression, least absolute shrinkage and selection operator (LASSO) regression, and multivariate Cox regression were used to further select biomarkers and construct a ΨrLs signature risk model. Then, the expression of lncRNAs of ΨrLs signature in multiple glioma cell lines was detected by qPCR. Further, ROC analysis, stratification analysis, correlation analysis, survival analysis, nomogram, enrichment analysis, immune infiltration analysis, chemoradiotherapy sensitivity analysis, somatic mutation, and recurrent copy number variation (CNV) analysis were used to validate the predictive efficiency of ΨrLs signature in TCGA and CGGA datasets.RESULTS: A four-lncRNA ΨrLs signature (DNAJC27-AS1, GDNF-AS1, ZBTB20-AS4, and DNMBP-AS1) risk model was constructed. By ROC analysis, stratified analysis, correlation analysis, survival analysis, and nomogram, the signature showed satisfactory predictive efficiency. Functional enrichment analysis revealed the differences in immune-related biological processes between high- and low-risk groups. Immune infiltration analysis showed that the high-risk group had lower tumor purity and higher stromal, immune and ESTIMATE scores. Mitoxantrone was identified as effective drug for low-risk group of glioma patients. Key genes in glioma development, including IDH1, EGFR, PTEN, etc., were differentially mutated between risk groups. The main recurrent CNVs in low-risk groups were 19q13.42 deletion and 7q34 amplification; 10q23.31 deletion and 12q14.1 in the high-risk group.CONCLUSIONS: Our study identified a four-lncRNA ΨrLs signature that effectively predicts the prognosis of glioma patients and may serve as a diagnostic tool. Risk scores of glioma patients generated by the signature is associated with immune-related biological processes and chemoradiotherapy sensitivity. These findings may inform the development of more targeted and effective therapies for glioma patients.
engineering, biomedical,computer science, interdisciplinary applications,mathematical & computational biology,biology