Methods for collapsing multiple rare variants in whole-genome sequence data.

Area of Science:

• Genomics and Statistical Genetics
• Cardiovascular Disease Research

Background:

• Whole-genome sequencing generates vast amounts of rare genetic variants.
• Hypertension and related traits require robust genetic analysis methods.
• Evaluating methods for rare variant analysis is crucial for genetic discovery.

Purpose of the Study:

• To evaluate the performance of various statistical methods for rare variant association analysis.
• To assess different collapsing strategies (gene-based vs. sliding windows) for rare variants.
• To identify challenges in analyzing rare variants from whole-genome sequence data in a pedigree sample.

Main Methods:

• Nine contributions evaluated collapsing methods using whole-genome sequence and longitudinal phenotype data.
• Methods included burden tests, variance-components tests, hybrid approaches, and novel techniques like functional principal components analysis.
• Genomic regions analyzed were gene-based or nonoverlapping sliding windows.

Main Results:

• Method performance varied based on genomic region characteristics, effect size, and variant direction.
• Most statistical methods demonstrated poor power to detect rare causal variants, performing similarly to type I error rates.
• Exceptions included MAP4 and FLT3; overall poor performance was attributed to small sample size, small variant effects, incomplete annotation, and linkage disequilibrium.

Conclusions:

• Analyzing rare variants from whole-genome sequence data presents significant challenges.
• Current statistical methods often lack sufficient power for rare variant detection in complex diseases like hypertension.
• Improvements in sample size, variant annotation, and analytical approaches are needed for effective rare variant analysis.