Stable gene expression for normalisation and single-sample scoring.

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Gene expression signatures are crucial for defining molecular phenotypes and classifying breast cancer subtypes (PAM50).
• Current methods like whole-transcriptome sequencing are costly and require substantial fresh tissue, limiting the clinical use of many gene signatures.
• Existing gene signature repositories contain thousands of valuable molecular profiles.

Purpose of the Study:

• To develop a novel method for quantifying gene expression signatures using stably expressed genes.
• To enable the clinical translation of gene expression signatures through cost-effective targeted measurements.
• To validate the stability and utility of the proposed method in cancer and normal tissues.

Main Methods:

• Identification of genes with stable expression and preserved relative ordering across diverse samples.
• Development of the 'stingscore' method to quantify signature gene expression using stably expressed genes.
• Validation of gene stability and signature score performance across various cancer and normal tissue datasets.

Main Results:

• A novel list of stably expressed genes was identified, demonstrating superior stability compared to existing gene sets.
• The stingscore method effectively quantifies relative expression levels of signature genes.
• Signature scores derived from targeted transcript measurements using stingscore accurately predicted docetaxel response in breast cancer patients.

Conclusions:

• The stingscore method offers a robust and stable approach for gene expression signature analysis.
• This method facilitates the development of targeted gene expression tests, supporting clinical applications.
• Gene expression signature analysis using stingscore can advance personalized medicine in oncology.