Bayesian Approach for Marker Selection in Genetic Studies
Author Information
Author(s): Yoon Seungtai, Suh Young Ju, Mendell Nancy Role, Ye Kenny Qian
Primary Institution: State University of New York at Stony Brook
Hypothesis
Can a Bayesian method improve the selection of genetic markers associated with phenotypes?
Conclusion
The study demonstrates a simpler Bayesian method for selecting important genetic markers that captures both main and interaction effects without extensive prior tuning.
Supporting Evidence
- Markers D3S4542 and sex showed high marginal posterior probabilities, indicating strong association with ALDX1.
- The method simplifies Bayesian variable selection and avoids the need for fine-tuning prior settings.
- The approach allows for flexibility in incorporating hypotheses into the model space.
Takeaway
This study shows a new way to find important genes that might affect health, using a method that is easier to use than older ones.
Methodology
The study analyzed microsatellite genotype data from the Genetics of Alcoholism study using a Bayesian variable selection method with a Metropolis-Hasting algorithm.
Limitations
The study does not address formal statistical inferences, which will be covered in future work.
Participant Demographics
The study included individuals classified as 'purely unaffected', 'unaffected with some symptoms', and 'affected' based on the ALDX1 variable.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website