Detecting heritable phenotypes without a model using fast permutation testing for heritability and set-tests.

Area of Science:

• Genetics
• Bioinformatics
• Statistical Genomics

Background:

• Genetic association studies are crucial for understanding phenotype heritability.
• Standard parametric methods for heritability and marker set testing rely on specific assumptions about phenotypic variability.
• These assumptions can lead to inflated p-values and false positives in analyses of methylation and gene expression data.

Purpose of the Study:

• To address the issue of inflated p-values in genetic association studies.
• To propose a novel method, FEATHER (Fast Estimation by permutation-based Association Testing using a HERitability framework), for robust heritability and marker set testing.
• To ensure proper control of false-positive results in high-throughput genetic analyses.

Main Methods:

• Developed FEATHER, a fast permutation-based testing method for marker sets and heritability.
• Applied FEATHER to heritability studies of methylation measurements and gene expression profiles.
• Compared FEATHER's results against standard parametric approaches.

Main Results:

• Parametric methods showed p-value inflation up to 15 orders of magnitude in methylation and gene expression studies.
• FEATHER effectively controlled for false-positive results.
• FEATHER identified 47% of methylation sites as non-heritable, which were previously flagged by parametric tests, indicating substantial false positives from alternative methods.

Conclusions:

• FEATHER provides a reliable and efficient approach for identifying heritable phenotypes from large-scale, high-throughput data.
• The method overcomes limitations of model misspecification inherent in parametric tests.
• FEATHER ensures accurate control of false-positive rates in genetic association studies.