Related Experiment Videos
Interaction screening by Kendall's partial correlation for ultrahigh-dimensional data with survival trait.
1Department of Statistics, Feng Chia University, Taichung 40724, Taiwan.
Bioinformatics (Oxford, England)
|January 12, 2020
Summary
We developed a new statistical method to identify gene interactions affecting survival outcomes, even with complex data. This approach improves the detection of important genetic markers for diseases like lung cancer.
Area of Science:
- Genomics
- Biostatistics
- Computational Biology
Background:
- Identifying gene interactions is crucial for understanding complex diseases.
- Ultrahigh-dimensional data and censored survival outcomes present significant challenges in genetic association studies.
- Existing methods struggle with contaminated data and interaction effect screening.
Purpose of the Study:
- To develop a robust statistical method for identifying interaction effects in genome-wide association studies (GWAS) with survival data.
- To address challenges posed by ultrahigh-dimensional data, contaminated datasets, and right-censored survival outcomes.
- To improve the screening of both main and interaction genetic effects associated with survival traits.
Main Methods:
- Proposed an inverse probability-of-censoring weighted (IPCW) Kendall's tau statistic to assess biomarker associations with survival traits.
- Developed a Kendall's partial correlation statistic for measuring survival trait relationships with interaction variables, conditional on main effects.
- Utilized the Kendall's partial correlation for interaction screening in ultrahigh-dimensional settings.
Main Results:
- Simulation studies demonstrated the proposed method's superior performance compared to existing approaches under various scenarios.
- Real-world data analysis successfully identified epistasis associated with clinical survival outcomes in non-small-cell lung cancer, diffuse large B-cell lymphoma, and lung adenocarcinoma.
- The method effectively identified both main and interaction biomarkers, outperforming current techniques.
Conclusions:
- The proposed IPCW-based method provides a powerful tool for interaction screening in GWAS with survival data.
- The method is robust to data contamination and censoring, offering improved accuracy in biomarker identification.
- The R-package 'IPCWK' is available for practical implementation, facilitating its use in genetic research.