Detecting Disease-Causing Genes Using LASSO-Patternsearch Algorithm
Author Information
Author(s): Shi Weiliang, Lee Kristine E, Wahba Grace
Primary Institution: University of Wisconsin Madison
Hypothesis
Can the LASSO-Patternsearch algorithm effectively identify disease-causing genes in rheumatoid arthritis data?
Conclusion
The modified LASSO-Patternsearch algorithm successfully identified important SNPs and covariates, achieving competitive error rates in classifying cases from controls.
Supporting Evidence
- The algorithm identified 145 main effects and 1439 interactions in the screening step.
- The final model included only 6 main effects and no interactions.
- The prediction error rate was 13.8% with a sensitivity of 85.3% and specificity of 87.0%.
Takeaway
This study created a new method to find genes that cause rheumatoid arthritis by looking at a lot of genetic information and figuring out which parts are important.
Methodology
The study used a modified LASSO-Patternsearch algorithm on simulated rheumatoid arthritis data to identify SNPs and covariates.
Limitations
Some important loci were missed due to their location on chromosomes and the algorithm's conservative nature.
Participant Demographics
The study involved simulated data from 1500 families with affected sib pairs and 2000 control families.
Statistical Information
P-Value
10-15
Statistical Significance
p<0.05
Want to read the original?
Access the complete publication on the publisher's website