locfdr

locfdr computes local false discovery rates and implements conditional false discovery rate methods in R to improve detection of genetic associations in genome-wide association studies by incorporating auxiliary continuous covariates.

Key Features:

• R implementation: Implemented as an R package providing functions for local FDR and conditional FDR (cFDR) calculations.
• Local FDR computation: Provides methodologies for calculating local false discovery rates to distinguish true signals from noise in large genomic datasets.
• Flexible conditional FDR (cFDR): Extends the traditional FDR framework by conditioning on auxiliary data to call significant associations.
• Support for arbitrary continuous distributions: Accepts auxiliary covariates from arbitrary continuous distributions rather than requiring specific parametric assumptions such as GWAS p-values.
• Iterative application for multi-dimensional data: Can be applied iteratively to accommodate multi-dimensional covariate data while maintaining control over the false discovery rate.
• Increased sensitivity with controlled FDR: Demonstrates increased sensitivity for detecting true associations without compromising FDR control as shown in simulation studies.

Scientific Applications:

• Genome-wide association studies (GWAS): Enhances identification of significant genetic variants by integrating auxiliary covariates, leveraging pleiotropy and non-random SNP distributions.
• Functional genomic data integration: Enables incorporation of diverse functional genomic datasets and test statistics from related traits to uncover additional genetic associations.
• Validation in large independent cohorts: Has been applied to an asthma GWAS with follow-up validation using the UK Biobank to identify and validate novel associations.

Methodology:

Implemented in R, the methods explicitly include computation of local FDR, conditional FDR conditioning on auxiliary covariates including arbitrary continuous distributions, iterative application for multi-dimensional covariate integration, and evaluation via simulation studies.