Associating Multivariate Quantitative Phenotypes with Genetic Variants in Family Samples with a Novel Kernel Machine...

Area of Science:

• Genetics
• Biostatistics
• Computational Biology

Background:

• Advanced sequencing technologies enable comprehensive analysis of genetic variants linked to complex diseases.
• Existing rare variant association tests require careful handling of familial correlation to prevent inflated type I error rates in family-based studies.

Purpose of the Study:

• To propose a novel multivariate family kernel machine regression (MF-KM) approach for analyzing genetic associations with multiple correlated phenotypes in family data.
• To ensure appropriate handling of familial correlation to maintain accurate type I error rates.

Main Methods:

• Developed MF-KM based on a linear mixed-model framework, suitable for diverse trait types.
• Compared MF-KM performance against standard kernel machine tests and single-phenotype analyses using simulations.
• Applied the methodology to whole-genome genotyping data from a lung function study.

Main Results:

• The standard kernel machine test showed inflated type I error rates on familial data, whereas MF-KM maintained expected rates.
• MF-KM demonstrated increased statistical power compared to methods analyzing phenotypes separately or using only unrelated individuals.
• The method was successfully illustrated using real-world lung function study data.

Conclusions:

• MF-KM offers a robust and powerful approach for multivariate genetic association studies in families.
• This method effectively addresses the challenges of familial correlation, improving gene detection for complex diseases.
• MF-KM provides a valuable tool for genetic research utilizing family-based whole-genome data.