An AI-Driven Multi-Omics Framework Identifies CASP8 as a Clinically Actionable Pyroptosis Biomarker in Bladder...

Area of Science:

• Oncology
• Genomics
• Computational Biology

Background:

• Translating multi-omics data into clinical biomarkers for bladder cancer is hindered by the complexity of genomic instability.
• Existing methods struggle to link genomic alterations to actionable biological insights for bladder cancer treatment.

Purpose of the Study:

• To develop an AI-driven framework for multi-omics data integration to identify bladder cancer biomarkers.
• To establish a link between genomic instability, pyroptosis, and clinical outcomes in bladder cancer.
• To identify and validate CASP8 as a clinically relevant biomarker for bladder cancer.

Main Methods:

• Integrated single-cell RNA sequencing, multi-cohort transcriptomics, and machine learning for genomic inference.
• Analyzed chromosomal aneuploidy and copy number variations at single-cell resolution to identify malignant cell populations.
• Developed a pyroptosis scoring system and employed machine learning for biomarker screening and experimental validation.

Main Results:

• Identified malignant cell populations and a consensus pyroptosis scoring system.
• Found elevated pyroptosis in the bladder cancer microenvironment, but suppressed pyroptosis in malignant cells with high genomic instability.
• Discovered low CASP8 expression correlates with high tumor mutation burden, driver gene alterations (TP53, RB1), and poor survival.

Conclusions:

• CASP8 is a clinically actionable biomarker linking genomic instability and pyroptosis dysregulation in bladder cancer.
• AI-integrated multi-omics analysis provides a powerful framework for biomarker discovery and translation.
• CASP8 loss promotes malignant phenotypes and alters cell death programs, impacting bladder cancer progression.