Identifying Causal Gene-Tissue Pairs and Variants Using TGVIS
Author Information
Author(s): Yang Yihe, Lorincz-Comi Noah, Zhu Xiaofeng
Primary Institution: Case Western Reserve University
Hypothesis
Can a multivariable TWAS method effectively identify causal gene-tissue pairs and direct causal variants while accounting for infinitesimal effects?
Conclusion
The TGVIS method improves the identification of causal gene-tissue pairs and variants, enhancing biological interpretability over existing methods.
Supporting Evidence
- TGVIS demonstrated lower mean square error in causal effect estimates compared to existing methods.
- TGVIS identified a higher proportion of colocalized credible sets than TGFM.
- The method improved precision in identifying causal genes related to lipid traits.
- TGVIS reduced the number of false positives by modeling infinitesimal effects.
- Significant differences in functionality were observed between direct causal variants and xQTLs.
Takeaway
This study created a new method to find important genes and their effects on traits, helping scientists understand how genes work together in different tissues.
Methodology
The study used a multivariable TWAS method called TGVIS, which incorporates infinitesimal effects and employs SuSiE for fine-mapping.
Potential Biases
The reliance on bulk tissue expression data may lead to misidentification of causal genes due to shared eQTLs.
Limitations
The analysis was limited to cis-eQTL/sQTL summary statistics, potentially missing trans-eQTL effects and specific cell-type causal genes.
Participant Demographics
The study analyzed data from 45 cardiometabolic traits across various tissues.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website