Explaining Gene Expression Using Twenty-One MicroRNAs.

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• The tumor transcriptome holds crucial disease information, but its regulation, particularly by microRNAs (miRs), remains incompletely understood.
• MicroRNAs are known to be involved in cancer development, highlighting the need for deeper investigation into their role in tumorigenesis.

Purpose of the Study:

• To investigate the relationship between microRNAs and various cancer types.
• To identify clusters of miRs associated with specific cancers and their contribution to gene expression regulation.

Main Methods:

• Analysis of approximately 9000 tumor samples from 32 cancer types within The Cancer Genome Atlas (TCGA).
• Application of a feature reduction algorithm to identify biologically interpretable miR clusters.
• Utilized linear models to quantify the variation in gene expression explained by miR cluster scores and tissue of origin.

Main Results:

• Identified 21 distinct miR clusters, many significantly associated with specific cancer types, capable of distinguishing between most cancers.
• Mean differences between tissues of origin explained 36% of gene expression variation; 21 miR cluster scores explained 30%.
• Combining tissue type and miR cluster scores explained 56% of the total genome-wide gene expression variation.

Conclusions:

• MicroRNA clusters are significant regulators of the cancer transcriptome, contributing substantially to gene expression variation.
• Genes involved in internal cellular processes (transport, metabolism) are more regulated by miRs than those involved in external signal reception (olfactory, sensory).
• Transcription factors and methylation may account for remaining gene expression variation, suggesting a multi-layered regulatory network in cancer.