Predicting master transcription factors from pan-cancer expression data

Area of Science:

• Oncology
• Genomics
• Molecular Biology

Background:

• Master transcription factors (MTFs) are crucial for development but can be hijacked during cancer to drive oncogenic programs.
• Current methods for identifying MTFs, primarily ChIP-seq, are limited by data availability across diverse cancer types.

Purpose of the Study:

• To develop a novel algorithm, Cancer Core Transcription factor Specificity (CaCTS), for identifying candidate MTFs using publicly available pan-cancer RNA sequencing data.
• To predict MTFs in cancer types and subtypes where they are currently unknown, expanding the scope of potential therapeutic targets.

Main Methods:

• Developed the CaCTS algorithm to analyze pan-cancer RNA sequencing data and prioritize candidate MTFs.
• Applied CaCTS across 34 tumor types and 140 subtypes to identify potential MTFs.
• Validated predicted MTFs (PAX8, SOX17, MECOM) in ovarian cancer (OvCa) models, assessing their role in cell viability, genomic localization, and response to transcriptional inhibition.

Main Results:

• CaCTS successfully identified candidate MTFs across a wide range of cancer types and subtypes, including novel predictions for previously uncharacterized cancers.
• Predicted MTFs, such as PAX8, SOX17, and MECOM in ovarian cancer, were confirmed to be essential for cancer cell viability.
• Validated MTFs were found to be proximal to superenhancers, co-occupy regulatory elements, bind loci encoding cancer biomarkers, and are sensitive to transcription-inhibiting drugs.

Conclusions:

• The CaCTS algorithm provides a robust method for predicting master transcription factors using RNA sequencing data, overcoming limitations of previous approaches.
• These findings expand the understanding of transcriptional drivers in numerous cancers, particularly those with limited existing knowledge.
• The identification and validation of novel MTFs offer promising new avenues for targeted cancer therapies across a broad spectrum of malignancies.