Association mapping via regularized regression analysis of single-nucleotide-polymorphism haplotypes in...

Area of Science:

• Genetics
• Statistical Genomics
• Bioinformatics

Background:

• Large-scale haplotype association analysis, particularly whole-genome scans, presents significant computational and statistical challenges.
• Current methods struggle with the complexity and scale of analyzing numerous haplotypes across diverse linkage disequilibrium patterns.

Purpose of the Study:

• To develop an efficient and effective method for large-scale haplotype association analysis.
• To address the challenge of multiple degrees of freedom in haplotype testing.
• To provide a framework that can integrate environmental risk factors into genetic association studies.

Main Methods:

• A variable-sized sliding-window framework is employed.
• Regularized regression analysis is utilized to manage multiple degrees of freedom.
• Window size is dynamically determined by local haplotype diversity and sample size.

Main Results:

• The proposed method demonstrates superior performance compared to single-nucleotide polymorphism (SNP) tests, cladistic haplotype analysis, and variable-length Markov chains.
• Outperformance is particularly notable in regions with high haplotype diversity.
• The method shows consistent results across various simulated disease models.

Conclusions:

• The novel sliding-window approach offers a more efficient and effective solution for large-scale haplotype association studies.
• The framework's ability to adapt window size makes it suitable for whole-genome scans with varying linkage disequilibrium.
• The integration of risk factors enhances its utility for complex disease research.