Uncovering Latent Structure in Gliomas Using Multi-Omics Factor Analysis
Catarina Gameiro Carvalho1, Alexandra M Carvalho2, Susana Vinga3,4
1Instituto Superior Técnico, Universidade de Lisboa, 1049-001 Lisbon, Portugal.
Genes
|May 27, 2026
Summary
This study used multi-omics analysis to find molecular differences in glioma brain tumors. The findings reveal new biomarkers and gene expression patterns that could lead to personalized cancer treatments.
Area of Science:
- Neuro-oncology
- Genomics
- Molecular Biology
Background:
- Gliomas are common adult malignant brain tumors with poor prognosis.
- Current World Health Organization (WHO) classification has limitations due to tumor heterogeneity.
- Advancements in sequencing and The Cancer Genome Atlas (TCGA) enable multi-level molecular investigation.
Purpose of the Study:
- To apply integrative multi-omics analysis to gliomas.
- To explore the interplay between genomic, epigenomic, and transcriptomic data.
- To identify molecular profiles and potential therapeutic targets in glioma.
Main Methods:
- Integrative multi-omics analysis combining genomic, epigenomic (DNA methylation), and transcriptomic (mRNA, miRNA) data.
- Utilized Multi-Omics Factor Analysis (MOFA), a Bayesian latent factor model.
- Analyzed data from The Cancer Genome Atlas (TCGA).
Main Results:
- Distinct molecular profiles were identified across oligodendroglioma, astrocytoma, and glioblastoma.
- Potential relationships between DNA methylation and gene expression were uncovered.
- Novel candidate biomarkers associated with survival and a transcriptional profile linked to neural development were discovered.
Conclusions:
- Findings highlight distinct glioma molecular subtypes.
- Identified potential biomarkers and gene expression patterns for survival.
- Results may inform personalized therapeutic strategies for improved treatment and outcomes.
