Mixed-effects models for GAW18 longitudinal blood pressure data.

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Longitudinal blood pressure data analysis requires sophisticated statistical models to account for repeated measurements and genetic factors.
• Existing methods like EMMA (Efficient Mixed-Model Association-mapping) correct for population structure but may not fully capture complex dependence structures in longitudinal data.

Purpose of the Study:

• To propose and evaluate two novel mixed-effects models for the analysis of longitudinal blood pressure data from the Genetic Analysis Workshop 18 (GAW18).
• To identify single nucleotide polymorphisms (SNPs) associated with systolic blood pressure (SBP) and diastolic blood pressure (DBP) using these new models.

Main Methods:

• An extension of the EMMA algorithm incorporating an estimated correlation matrix to model the dependence structure of repeated measurements.
• A Bayesian multiple association-mapping algorithm with variable selection, capable of modeling multiple SNPs, SNP-SNP, and SNP-environment interactions.

Main Results:

• The extended EMMA method identified significant SNPs for SBP and DBP, including Chr5:75506197 for SBP and SNPs on Chr3, Chr17, and Chr21 for DBP.
• The Bayesian method identified additional SNPs for SBP, with strong evidence (Bayes factors) for associations on Chr1, Chr3, Chr15, and Chr19.
• A significant SNP (Chr3:197469358) was found to interact with age for SBP in the Bayesian analysis.

Conclusions:

• The developed mixed-effects models are effective for identifying genetic associations in longitudinal blood pressure data.
• These methods offer advancements in handling complex data structures and identifying multiple genetic effects and interactions.