Multiomics and machine learning-based analysis of pancancer pseudouridine modifications

Area of Science:

• Oncology
• Molecular Biology
• Bioinformatics

Background:

• Pseudouridine is crucial for RNA stability and function, but its role in cancer remains unclear.
• Pseudouridine synthases (PUSs) are responsible for pseudouridine modification.
• Understanding PUSs in tumors is vital for cancer research.

Purpose of the Study:

• To systematically analyze PUSs expression, genomic aberrations, and prognostic value across 33 tumor types.
• To investigate the association of PUSs with signaling pathways and the tumor microenvironment (TME).
• To develop a prognostic model for guiding cancer treatment and immunotherapy.

Main Methods:

• Analysis of PUSs expression, copy number variations (CNVs), and DNA methylation in 10907 tumor samples.
• Functional enrichment analyses to identify associated signaling pathways (MYC, E2F, MTORC1).
• Machine learning models (over 112 combinations) to build a prognostic signature.

Main Results:

• PUSs pathway dysregulation was observed in tumors, affecting patient prognosis.
• PUSs expression is linked to CNVs, DNA methylation, and key signaling pathways.
• PUSs influence the TME, correlating with immune checkpoints and Treg infiltration in lung cancer.
• A validated prognostic signature showed efficacy in multiple tumor types and predicted treatment response.

Conclusions:

• PUSs are dysregulated in cancer, influencing the TME and patient outcomes.
• The developed PUSs-based model offers prognostic value and potential for guiding immunotherapy.
• Targeting pseudouridine modifications presents a promising avenue for novel cancer diagnostics and therapeutics.