LSMM: a statistical approach to integrating functional annotations with genome-wide association studies

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Genome-wide association studies (GWAS) have identified numerous risk variants for complex traits and diseases.
• Challenges remain in interpreting non-coding GWAS hits and identifying variants with small effects due to trait polygenicity.

Purpose of the Study:

• To develop a method that integrates functional annotation data with GWAS to improve the understanding of complex trait genetic architectures.
• To increase statistical power in identifying risk variants and gain biological insights into relevant functional annotations.

Main Methods:

• Proposed a latent sparse mixed model (LSMM) to integrate functional annotations with GWAS data.
• Developed an efficient variational expectation-maximization algorithm for scalable parameter estimation and statistical inference.
• Applied LSMM to analyze 30 GWAS datasets using genic and cell-type specific functional annotations.

Main Results:

• LSMM demonstrated increased statistical power for identifying risk variants compared to conventional methods.
• The model successfully detected relevant functional annotations, providing deeper biological insights.
• LSMM analysis of 30 GWAS datasets confirmed its effectiveness in understanding complex trait genetic architectures.

Conclusions:

• The developed LSMM method effectively integrates functional annotations with GWAS data.
• LSMM enhances the identification of genetic variants underlying complex phenotypes and offers greater biological interpretability.
• The LSMM software is publicly available for broader research application.