Classification models for clear cell renal carcinoma stage progression, based on tumor RNAseq expression trained
1Bioinformatics Laboratory, Structural and Computational Biology Group, International Centre for Genetic Engineering and Biotechnology (ICGEB), Aruna Asaf Ali Marg, New Delhi, India.
BMC Proceedings
|November 7, 2014
Summary
This study developed gene expression models to differentiate early and late-stage clear-cell Renal Cell Carcinoma (ccRCC). The Random Forest model showed the best performance, identifying key genes for understanding ccRCC progression.
Area of Science:
- Oncology
- Bioinformatics
- Genomics
Background:
- Clear-cell Renal Cell Carcinoma (ccRCC) is an aggressive kidney cancer with poor prognosis.
- Lack of early diagnostic and prognostic biomarkers hinders effective treatment for ccRCC.
- Molecular heterogeneity and asymptomatic early stages necessitate novel biomarker discovery.
Purpose of the Study:
- To develop and evaluate machine learning models for distinguishing early-stage from late-stage ccRCC.
- To identify a robust gene signature for ccRCC stage classification.
- To leverage The Cancer Genome Atlas (TCGA) gene expression data for ccRCC research.
Main Methods:
- Utilized supervised learning algorithms including J48, Random Forest, SMO, and Naïve Bayes.
- Employed Fast Correlation Based Feature Selection (FCBF) for feature enrichment.
- Trained and validated models using RNA sequencing gene expression data from TCGA.
Main Results:
- The Random Forest model demonstrated superior performance in ccRCC staging.
- Achieved 89% sensitivity, 77% accuracy, and an Area Under the ROC Curve (AUC) of 0.8.
- Identified a prioritized set of 62 genes with potential prognostic value.
Conclusions:
- The developed prediction models and 62-gene signature can aid in understanding ccRCC molecular mechanisms.
- These findings may accelerate the discovery of prognostic factors for ccRCC.
- The study provides a foundation for improved diagnostic and prognostic tools for ccRCC.
