Epigenetic Combinatorial Patterns Predict Disease Variants.

Area of Science:

• Genomics
• Computational Biology
• Human Genetics

Background:

• Genome-wide association studies (GWAS) identify numerous noncoding genetic variants, often tagging causal variants.
• Precisely locating disease-causal variants and understanding their function remains a significant challenge.
• Integrating functional data from diverse human tissues and cell types offers a promising approach for variant fine-mapping.

Purpose of the Study:

• To investigate if analyzing combinatorial patterns of functional data across cell types enhances disease variant prediction accuracy.
• To develop a new method for prioritizing disease-associated variants by leveraging cell-type-specific functional information.

Main Methods:

• A Bayesian method was developed to identify recurring cell-type-specificity partitions across the human genome using functional annotation from 127 cell types.
• Epigenetic cell-type specificity and functional element enrichment were utilized for variant prioritization.

Main Results:

• The de novo identified epigenome partition patterns aligned well with known cell-type origins.
• Associated functional elements showed significant enrichment in disease variants.
• The proposed method demonstrated substantially improved prediction power for disease variants compared to linear models.

Conclusions:

• Analyzing combinatorial patterns of functional genomic data across cell types offers a powerful strategy for disease variant fine-mapping.
• The developed Bayesian approach provides an effective new tool for prioritizing functional disease variants for experimental validation.