Linkage analysis of complex disorders with multiple phenotypic categories: simulation studies and application to...

Area of Science:

• Genetics
• Statistical genetics
• Bioinformatics

Background:

• Linkage analysis for disorders with multiple phenotypes (diagnostic spectrum) presents challenges in accurately identifying disease-associated genes.
• Existing methods, like Ott's [1994] down-weighting approach, may require refinement for complex genetic models.

Purpose of the Study:

• To propose a modified linkage analysis method that accounts for the diagnostic spectrum of disorders.
• To introduce a 'robust weighting' strategy for more reliable genetic model assessment.
• To evaluate the performance of the weighted model against existing methods using real-world data.

Main Methods:

• A modification of Ott's method is proposed, reducing penetrance ratios for broader diagnostic categories.
• A 'robust weighting' strategy assesses the consistency of genetic model ratios across various scenarios.
• An additive parametric analysis is compared with dominant, recessive, and nonparametric linkage (NPL) analyses.

Main Results:

• Additive parametric analysis showed high correlation with dominant, recessive, and NPL analyses.
• The weighted, additive model performed effectively on a modified NIMH bipolar chromosome 18 dataset.
• The weighted model demonstrated strong performance when compared to NPL analyses under different diagnostic models.

Conclusions:

• The proposed robust weighting strategy offers an improved approach for linkage analysis in disorders with phenotypic variability.
• Weighted parametric models correlate well with traditional linkage analysis methods.
• Incorporating similar weighting strategies into nonparametric linkage analyses could further enhance their utility.