Machine Learning-Enhanced Analysis of miRNA Biomarkers for Accurate Breast Cancer Diagnosis Using DNA Seagrass

Area of Science:

• Biochemistry
• Molecular Biology
• Bioinformatics

Background:

• MicroRNAs (miRNAs) show promise as breast cancer biomarkers.
• Current detection methods lack sensitivity, specificity, and simplicity.
• Reliable biomarkers are needed for accurate miRNA-based diagnosis.

Purpose of the Study:

• To develop an ultrasensitive and highly specific method for detecting multiple miRNA biomarkers for breast cancer diagnosis.
• To integrate machine learning for precise breast cancer diagnosis using miRNA panels.
• To advance clinical applications of miRNA detection in breast cancer.

Main Methods:

• Developed rolling circle amplification-generated DNA seaweed (RCA-GDS) for efficient miRNA signal amplification.
• Screened and validated a panel of miRNAs (miRNA21, miRNA182, miRNA183) using the TCGA database.
• Integrated machine learning algorithms with the RCA-GDS detection system for differential diagnosis.

Main Results:

• RCA-GDS achieved attomolar-level miRNA detection within 2 hours under isothermal conditions.
• Validated miRNA panel reliability in both intracellular and serum samples.
• The integrated machine learning model demonstrated excellent diagnostic accuracy in an independent cohort.

Conclusions:

• The RCA-GDS method enables ultrasensitive and highly specific miRNA detection.
• The combined approach advances miRNA panel-based breast cancer diagnosis.
• This strategy improves accuracy and clinical applicability for breast cancer detection.