Using Logistic Regression Trees to Select Genetic Markers
Author Information
Author(s): Nickolov Radoslav Z, Milanov Valentin B
Primary Institution: Department of Mathematics and Computer Science, Fayetteville State University
Hypothesis
Can the LOTUS algorithm effectively reduce a large set of genetic markers to a smaller set of interesting SNPs?
Conclusion
The LOTUS algorithm is effective in selecting a small number of interesting SNPs from a large pool in case-control studies.
Supporting Evidence
- Trait loci C and E were selected in the set of interesting SNPs in all 50 replications.
- Incorporating MDR significantly improved the selection rate of trait loci A, B, D, and F.
- The false-positive rates for the first procedure were close to 5%, while the second procedure nearly doubled this rate.
Takeaway
This study shows a way to pick out important genetic markers from a big list, making it easier to study diseases.
Methodology
The study used simulated data from Genetic Analysis Workshop 15, applying the LOTUS algorithm to select SNPs from a large set.
Potential Biases
The second selection procedure may increase the false-positive rate.
Limitations
The study only included genetic markers and may not generalize to other types of predictors.
Participant Demographics
The study used simulated data with 1000 cases and 1000 controls.
Statistical Information
P-Value
0.05
Statistical Significance
p<0.05
Want to read the original?
Access the complete publication on the publisher's website