Penalized regression approaches to testing for quantitative trait-rare variant association

Area of Science:

• Statistical data analysis
• Genetics
• Bioinformatics

Background:

• Penalized regression offers simultaneous variable selection and parameter estimation.
• Existing penalized regression methods excel in high-dimensional data but are underutilized in hypothesis testing, particularly for genetic association analysis.
• There is a need for advanced methods to analyze associations between quantitative traits and groups of rare variants.

Purpose of the Study:

• To apply novel penalized regression methods with a Truncated L1-penalty (TLP) for variable selection and parameter grouping in genetic association analysis.
• To evaluate the performance of these new methods against existing penalized regression and global association tests.
• To explore the application of penalized regression in hypothesis testing for rare variant association studies.

Main Methods:

• Application of several new penalized regression methods incorporating the Truncated L1-penalty (TLP).
• Data-adaptive variable selection and parameter grouping strategies were employed.
• Performance evaluation through simulations and analysis of real sequence data from the Genetic Analysis Workshop 17 (GAW17).
• Comparison with existing penalized regression and global association tests.

Main Results:

• The proposed penalized methods showed improvements over some existing penalized approaches.
• However, the new methods did not consistently outperform established global association tests.
• The study identified potential challenges in applying penalized regression methods to genetic hypothesis testing.

Conclusions:

• Penalized regression methods, including those with TLP, show promise for selecting causal variants in genetic association studies.
• Further research is warranted to refine these methods and address limitations in genetic hypothesis testing.
• The application of penalized regression in analyzing rare variant associations requires careful consideration and further investigation.