On the interpretation of transcriptome-wide association studies

Area of Science:

• Genetics
• Bioinformatics
• Statistical Genomics

Background:

• Transcriptome-wide association studies (TWAS) are widely used for analyzing genome-wide association study (GWAS) data.
• TWAS aims to identify relationships between gene expression and specific phenotypes.
• Current interpretations of TWAS results often imply a genetic link, which may not align with the tested null hypothesis.

Purpose of the Study:

• To mathematically outline the TWAS framework and clarify its tested null hypothesis.
• To assess the implications of misinterpreting TWAS results as direct evidence of genetic relationships.
• To evaluate the reliability of TWAS for investigating gene expression-phenotype genetic links.

Main Methods:

• Mathematical formulation of the TWAS framework.
• Simulations to evaluate type 1 error rates under misinterpretation.
• Real data analysis of TWAS results.

Main Results:

• TWAS exhibits considerably inflated type 1 error rates when interpreted as indicative of genetic relationships.
• 41% of significant TWAS associations lacked sufficient statistical evidence for a genetic relationship.
• Misinterpretation leads to unreliable inferences about gene expression-phenotype genetic links.

Conclusions:

• Current TWAS implementations are unreliable for inferring genetic relationships between gene expression and phenotypes.
• Local genetic correlation analysis presents a potential alternative for robust investigation.
• Revising the interpretation of TWAS results is crucial for accurate genetic association studies.