Analyzing Gene Interactions with Rule-Based Classifiers
Author Information
Author(s): Thorsten Lehr, Jing Yuan, Dirk Zeumer, Supriya Jayadev, Marylyn D Ritchie
Primary Institution: Boehringer Ingelheim Pharma GmbH & Co. KG
Hypothesis
The study aims to evaluate the performance of three rule-based classifier algorithms for analyzing genetic association studies.
Conclusion
The study confirms that rule-based classifiers like RIPPER, RIDOR, and PART are useful for detecting gene-gene and gene-environment interactions in genetic association studies.
Supporting Evidence
- The RIPPER algorithm discovered the true case-control model in over 33% of the datasets.
- RIPPER, RIDOR, and PART algorithms detected true case-control rules in more than 83% of the datasets.
- All three algorithms were able to detect attributes used in the case-control models in most datasets.
Takeaway
This study tested different computer programs to see how well they can find connections between genes and diseases. One program, RIPPER, did the best job.
Methodology
The study simulated 42 datasets with varying subjects, SNPs, and noise, applying three rule-based classifiers to evaluate their performance.
Limitations
The study was limited to simulated datasets and did not test the algorithms on real-world data.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website