Application of collapsing methods for continuous traits to the Genetic Analysis Workshop 17 exome sequence data.

Area of Science:

• Genetics
• Bioinformatics
• Statistical Genetics

Background:

• Rare variants play a significant role in complex diseases.
• Accurate identification of disease-associated rare variants is challenging.
• Existing methods for analyzing multiple rare variants require evaluation.

Purpose of the Study:

• To assess the performance of various rare variant collapsing methods.
• To compare collapsing methods against single-marker analysis.
• To evaluate methods using real sequence data and simulated phenotypes.

Main Methods:

• Applied collapsing methods to continuous phenotypes (Q1, Q2) using 1000 Genomes Project data.
• Collapsed SNPs based on MAF and coding status (all, MAF<0.05, nonsynonymous MAF<0.05).
• Compared collapsing methods (proportion vs. presence/absence tests) with single-marker analysis (PLINK).

Main Results:

• Collapsing rare nonsynonymous SNPs using a proportion test performed best for phenotype Q1, identifying genes FLT1 and KDR.
• Single-marker analysis also yielded significant results for SNPs in FLT1 and KDR.
• For phenotype Q2, collapsing rare nonsynonymous SNPs was effective, but no methods reached statistical significance at true causal genes.
• High correlations between noncausal and causal genes may inflate false positive rates.

Conclusions:

• Rare variant collapsing methods, especially for nonsynonymous variants, show potential in genetic association studies.
• Challenges persist in detecting true causal variants, particularly for complex phenotypes.
• Confounding effects from correlated noncausal genes necessitate careful interpretation of results.