PALM: a powerful and adaptive latent model for prioritizing risk variants with functional annotations

Area of Science:

• Genomics
• Statistical Genetics
• Bioinformatics

Background:

• Genome-wide association studies (GWAS) have advanced understanding of complex traits but face challenges.
• Many GWAS hits are in non-coding regions with unknown functions.
• Polygenicity hinders identification of weak-to-moderate effect risk variants.

Purpose of the Study:

• To develop a novel method for integrating cell-type/tissue-specific functional annotations with GWAS summary statistics.
• To address limitations of existing linear models in characterizing complex genetic architectures.
• To improve the identification and interpretation of genetic risk variants for human complex diseases.

Main Methods:

• Proposed a powerful and adaptive latent model (PALM) utilizing a tree ensemble.
• Developed a functional gradient-based expectation-maximization algorithm for scalable model fitting.
• Integrated cell-type/tissue-specific functional annotations with GWAS summary statistics.

Main Results:

• PALM effectively controls the false discovery rate and enhances statistical power for identifying risk variants.
• The model adaptively characterizes non-linear relationships between functional annotations and variant association.
• Application to 30 GWAS datasets identified more risk variants and ranked functional annotation importance.

Conclusions:

• PALM offers a powerful and adaptive approach for integrating functional genomics data with GWAS.
• The method improves the discovery of genetic risk variants and aids in interpreting GWAS findings.
• PALM provides a scalable and effective tool for advancing genetic research in complex diseases.