Pan-Cancer Molecular Patterns and Biological Implications Associated with a Tumor-Specific Molecular Signature

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Defining pan-cancer subtypes requires robust molecular signatures.
• The PAM50 signature is FDA-approved for breast cancer classification.
• Controlling classification errors is crucial for tissue-independent cancer analysis.

Purpose of the Study:

• To repurpose the PAM50 signature for pan-cancer subtype discovery.
• To identify tumor-agnostic molecular patterns using PAM50 with uncertainty assessment.
• To explore the biological and clinical implications of identified pan-cancer subtypes.

Main Methods:

• Applied PAM50 signature with uncertainty assessment to over 33 cancer types.
• Discarded unassigned samples and analyzed remaining tumor-agnostic patterns.
• Utilized gene set analysis to characterize emerging molecular classes (C1 and C2).

Main Results:

• Identified two pan-cancer molecular classes (C1 and C2) from classified samples.
• Class C2 exhibited dedifferentiation, higher proliferation, centrosome amplification, and potential TP53/RB1 mutations.
• Found 28 gene sets and 95 genes linked to cell-cycle and DNA damage pathways in C2.
• Observed associations between C1/C2 classification, survival, drug sensitivity, and immune infiltrate prognostic value in some cancers.

Conclusions:

• PAM50, with uncertainty assessment, can be repurposed for pan-cancer classification.
• The identified C1 and C2 subtypes possess distinct molecular, biological, and clinical characteristics.
• This approach offers a novel framework for understanding tumor-agnostic cancer biology and treatment response.