Accounting for nonlinear effects of gene expression identifies additional associated genes in transcriptome-wide...

Area of Science:

• Genetics
• Bioinformatics
• Complex Trait Genetics

Background:

• Transcriptome-wide association studies (TWAS) integrate genome-wide association study (GWAS) and gene expression (GE) data to identify causal genes for complex traits.
• Standard TWAS assumes a linear relationship between GE and traits, potentially missing non-linear associations and reducing statistical power.

Purpose of the Study:

• To extend standard TWAS by incorporating a quadratic effect of gene expression alongside the linear effect.
• To improve the identification of genes associated with complex traits by accounting for potentially non-linear GE-trait relationships.

Main Methods:

• Developed a two-stage TWAS approach: Stage 1 trains imputation models for both linear and quadratic GE levels from genotypes.
• Stage 2 tests gene-trait association using both imputed linear and quadratic GE levels.
• Applied the enhanced TWAS method to Alzheimer's disease and lipid GWAS data, using ADNI, IGAP, GTEx, and UK Biobank datasets.

Main Results:

• The enhanced TWAS approach identified additional genes associated with Alzheimer's disease.
• The method also revealed new gene associations for LDL and HDL cholesterol levels.
• Demonstrated potential power gains compared to standard TWAS by capturing non-linear GE effects.

Conclusions:

• The inclusion of quadratic gene expression effects in TWAS can uncover additional trait-associated genes.
• This nonlinear TWAS approach offers improved power for identifying causal genes for complex traits.
• The findings highlight the importance of considering non-linear gene expression effects in genetic association studies.