Funmap: integrating high-dimensional functional annotations to improve fine-mapping.

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Fine-mapping prioritizes causal variants in complex traits using genome-wide association study (GWAS) data.
• Functional annotations enhance fine-mapping power but existing methods often yield false positives with large datasets.

Purpose of the Study:

• To develop a unified method (Funmap) for integrating high-dimensional functional annotations into fine-mapping.
• To improve fine-mapping power while maintaining a controlled false positive rate.

Main Methods:

• Funmap integrates hundreds of functional annotations using a random effects model to link annotations with causal probabilities.
• A fast algorithm enables scalable integration of numerous annotations for prioritizing multiple causal single nucleotide polymorphisms (SNPs).

Main Results:

• Simulations show Funmap controls false discovery rates under high-dimensional annotations, unlike other methods.
• Funmap achieves comparable or better power gains than existing methods.
• Application to 4 lipid GWAS datasets with 187 annotations revealed higher replication rates (15.5%-26.2% improvement).

Conclusions:

• Funmap offers a powerful and well-calibrated approach for fine-mapping complex traits using extensive functional annotations.
• The method enhances variant prioritization and improves replication rates in independent cohorts.