Rapid and robust association mapping of expression quantitative trait loci

Area of Science:

• Genetics
• Bioinformatics
• Statistical Genetics

Background:

• Gene expression quantitative trait loci (eQTL) analysis is crucial for understanding genotype-phenotype relationships.
• Mixed model frameworks are commonly used for family-based association studies of eQTLs.
• Computational efficiency and accurate detection of true positives are key challenges in eQTL analysis.

Purpose of the Study:

• To develop and evaluate a computationally efficient two-step method for analyzing family-based eQTL studies.
• To demonstrate the impact of data filtering strategies on reducing computational burden and controlling false positives.
• To compare the performance of the proposed method against a full mixed model approach.

Main Methods:

• Application of a two-step analysis method within a mixed model framework for family-based eQTL studies.
• Implementation of data filtering based on expression variability and genotype counts.
• Comparative analysis using the Genetic Analysis Workshop 15 (GAW15) Problem 1 dataset.
• Focusing analysis on cis-acting regulatory elements by considering proximity between markers and transcripts.

Main Results:

• The two-step method yielded results comparable to the full mixed model but was significantly faster.
• Filtering non-expressed genes based on expression variability reduced the number of tests by approximately 50%.
• Filtering on genotype counts effectively minimized spurious associations.
• Analysis of cis-acting signals revealed a five-fold increase in detected signals compared to genome-wide analysis.

Conclusions:

• A two-step analytical approach offers a computationally efficient alternative for family-based eQTL studies.
• Strategic data pre-filtering is essential for reducing false positives and increasing the power to detect true genetic associations.
• Partitioning data and focusing on cis-regulatory regions can enhance the discovery of genuine eQTL effects.