Using Alternating Decision Trees to Model Disease Traits
Author Information
Author(s): Liu Kuang-Yu, Lin Jennifer, Zhou Xiaobo, Wong Stephen TC
Primary Institution: Brigham and Women's Hospital, Harvard Medical School
Hypothesis
Can alternating decision trees improve the modeling of disease traits using genetic data?
Conclusion
The ADTrees method provides a more accurate representation of disease status and better detection of genetic linkages.
Supporting Evidence
- ADTrees detected significant evidence of linkage on chromosomes 1, 3, 5, and 9.
- The full-ADT approach performed better than BDT in extracting relevant phenotypes.
- All relevant phenotypes were extracted for four populations except one false negative.
- Iteration scores suggested subgroup structures consistent with latent traits.
Takeaway
This study used a special computer method to better understand how certain traits are linked to diseases by looking at genetic information.
Methodology
The study applied ADTrees to genetic data from four populations, using 10-fold cross validation to assess model performance.
Limitations
The study did not recover the full simulated model and had one false negative in the Karangar population.
Participant Demographics
The study involved four populations: Aipotu, Danacca, Karangar, and NYC.
Statistical Information
P-Value
p<0.01
Statistical Significance
p<0.01
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website