A unifying statistical framework to discover disease genes from GWASs.

Area of Science:

• Genetics
• Genomics
• Bioinformatics

Background:

• Genome-wide association studies (GWASs) identify genomic regions linked to complex traits.
• Identifying specific causal genes within these regions is a significant challenge due to multiple confounding factors.

Purpose of the Study:

• To develop a unified mathematical framework to accurately identify genes affected by causal genetic variants.
• To compute the probability of each gene being influenced by a causal variant for a given trait without prior assumptions on cell types or tissues.

Main Methods:

• Developed a novel unified mathematical framework integrating solutions to challenges in gene identification from GWAS data.
• Employed a probabilistic approach to calculate the likelihood of a gene being affected by a causal variant.
• Validated individual components and the combined framework using established genetic principles and data.

Main Results:

• The framework successfully computes the probability of genes being affected by causal variants across the genome for specific traits.
• When applied to large human disease GWASs, the method rediscovered a majority of known disease genes.
• Provided human genetics support for numerous genes with prior clinical/preclinical evidence and identified novel disease gene candidates.

Conclusions:

• The developed framework offers a robust method for pinpointing disease-associated genes from GWAS data.
• This approach enhances the discovery of novel disease genes and strengthens evidence for existing candidates.
• It provides a powerful tool for understanding the genetic architecture of complex human diseases.