Related Experiment Video
Updated: Apr 19, 2026
07:41
Performing Data Mining And Integrative Analysis Of Biomarker in Breast Cancer Using Multiple Publicly Accessible Databases
Published on: May 17, 2019
9.7K
Network-based biomarkers enhance classical approaches to prognostic gene expression signatures
BMC Systems Biology
|December 19, 2014
Summary
Network-based gene expression analysis methods offer complementary insights to traditional approaches for predicting patient outcomes. Combining these methods can improve classification accuracy by leveraging unique patient data subsets identified by each technique.
Area of Science:
- Bioinformatics
- Computational Biology
- Genomics
Background:
- Traditional gene expression analysis for patient outcome prediction relies on single-gene or gene-set methods.
- Emerging network-based approaches utilize gene interaction information, but their added value remains unclear.
- Comparing traditional and network-based methods is crucial for advancing prognostic biomarker development.
Purpose of the Study:
- To compare the performance of single-gene, gene-set, and network-based methods for predicting patient clinical outcomes.
- To investigate whether network-based approaches offer advantages over traditional signature modeling techniques.
- To explore the value of different network-based strategies, including informative gene identification and sub-network analysis.
Main Methods:
- Utilized gene expression microarray data from melanoma and ovarian cancer patients.
- Implemented and compared single-gene, gene-set, and two types of network-based methods (informative genes and sub-networks).
- Performed 100 rounds of 5-fold cross-validation with three classifiers and two protein-protein interaction networks.
Main Results:
- All methods (single-gene, gene-set, network-based) showed similar error rates in melanoma and ovarian cancer datasets.
- Patient-level analysis revealed that different methods correctly classified distinct patient subsets.
- The NetRank network-based feature selection method demonstrated the highest stability.
Conclusions:
- Network-based gene expression signature modeling offers value by capturing different patient sample subspaces.
- Combining various classification methods can enhance overall accuracy by leveraging method-specific strengths.
- In-depth patient-level analysis is essential for understanding the complementary value of different predictive approaches.
