Evaluation of Clinically Significant miRNAs Level by Machine Learning Approaches Utilizing Total Transcriptome Data.

Area of Science:

• Bioinformatics
• Molecular Biology
• Cancer Research

Background:

• Cancer occurrence and progression mechanisms are key research objectives.
• Omics data, especially transcriptomes, offer detailed insights into RNA expression and regulation.
• Understanding RNA interactions is crucial for cancer bioinformatics.

Purpose of the Study:

• To analyze correlations between specific microRNAs (miRNAs) and other RNAs across diverse cancer types.
• To investigate the dependence of miRNA expression on small nucleolar RNAs (snoRNAs) and long noncoding RNAs (lncRNAs).
• To build a comprehensive database for identifying RNA expression dependencies in various cancers.

Main Methods:

• Assembled a large-scale transcriptomic dataset from ~16,000 cancer patients across >160 cancer types.
• Utilized gradient boosting algorithms to identify correlations between four specific miRNAs and ~60,660 other RNAs.
• Focused on hsa-mir-21, hsa-let-7a-1, hsa-let-7b, and hsa-let-7i expression levels.

Main Results:

• Discovered significant dependencies between the expression of studied miRNAs and specific snoRNAs and lncRNAs.
• Confirmed that the identified RNA roles in cancer development are supported by prior experimental evidence.
• Established a foundation for identifying broader RNA expression dependencies in diverse cancers.

Conclusions:

• The expression of key miRNAs in cancer is influenced by other regulatory RNAs like snoRNAs and lncRNAs.
• The developed database serves as a valuable resource for future cancer research and biomarker discovery.
• This work facilitates the identification of novel, cancer-specific RNA alterations.