A workflow to study mechanistic indicators for driver gene prediction with Moonlight.

Area of Science:

• Bioinformatics
• Cancer Genomics
• Systems Biology

Background:

• Predicting driver genes is crucial for understanding cancer development and identifying therapeutic targets.
• Existing bioinformatics frameworks like Moonlight integrate gene expression with cancer hallmarks and regulatory networks.
• Mechanistic indicators, such as mutations in regulatory regions, are needed to confirm driver genes and link them to observed expression changes.

Purpose of the Study:

• To introduce Moonlight2, an enhanced bioinformatics framework for driver gene prediction.
• To incorporate a mutation-based mechanistic indicator as a second layer of evidence.
• To analyze mutations within cancer cohorts and classify them as driver or passenger mutations.

Main Methods:

• Moonlight2 integrates gene expression data with mutation analysis.
• It identifies oncogenic mediators and then filters them using mutation data.
• Driver and passenger mutations are classified, and oncogenic mediators with driver mutations are identified as final driver genes.

Main Results:

• Moonlight2 was applied to basal-like breast cancer, lung adenocarcinoma, and thyroid carcinoma datasets from The Cancer Genome Atlas.
• In basal-like breast cancer, four oncogenes and nine tumor suppressor genes with promoter driver mutations were identified.
• These findings suggest potential explanations for the deregulation of these identified driver genes.

Conclusions:

• Moonlight2 provides a robust framework for driver gene prediction by combining gene expression and mutation data.
• The mutation-based mechanistic indicator refines the identification of driver genes.
• This approach aids in understanding cancer biology and discovering new therapeutic strategies.