Related Experiment Video
Updated: Jul 25, 2025
07:41
Performing Data Mining And Integrative Analysis Of Biomarker in Breast Cancer Using Multiple Publicly Accessible Databases
Published on: May 17, 2019
9.0K
Identification of Novel Diagnostic and Prognostic Gene Signature Biomarkers for Breast Cancer Using Artificial
Zeenat Mirza1,2, Md Shahid Ansari3, Md Shahid Iqbal4
1King Fahd Medical Research Center, King Abdulaziz University, Jeddah 21589, Saudi Arabia.
Cancers
|June 28, 2023
Summary
Machine learning models identified novel gene signatures for breast cancer (BC) diagnosis and prognosis. These models, based on differentially expressed genes (DEGs), offer improved accuracy for identifying BC and predicting patient outcomes.
Area of Science:
- Genomics
- Bioinformatics
- Machine Learning in Oncology
Background:
- Breast cancer (BC) diagnosis relies on clinical and histopathological data, which can lack precision.
- Machine learning (ML) offers a powerful approach to enhance diagnostic accuracy by analyzing gene expression patterns.
Purpose of the Study:
- To develop and validate ML-based diagnostic and prognostic models for breast cancer using gene expression profiling.
- To identify novel gene signatures indicative of BC presence and patient survival.
Main Methods:
- Analysis of 701 BC samples from 11 GEO microarray datasets to identify differentially expressed genes (DEGs).
- Application of seven ML methods for gene selection and construction of diagnostic and prognostic models.
- Validation of identified gene signatures using qRT-PCR and additional ML algorithms (GBDT, XGBoost, AdaBoost, KNN, MLP).
Main Results:
- Identification of 355 DEGs and prediction of BC-associated pathways.
- Discovery of a nine-gene diagnostic signature (including COL10A, S100P, ADAMTS5) and an eight-gene prognostic signature (including CCNE2, NUSAP1, TPX2).
- Validation of gene signatures through qRT-PCR and multiple ML methods, confirming their expression and predictive potential.
Conclusions:
- ML approaches successfully constructed novel diagnostic and prognostic models for breast cancer.
- The identified nine-gene diagnostic and eight-gene prognostic signatures demonstrate significant potential for improving BC detection and outcome prediction.
