Comparing Clustering Methods for Genetic Data
Author Information
Author(s): Tricia A Thornton-Wells, Jason H Moore, Jonathan L Haines
Primary Institution: Vanderbilt University Medical Center
Hypothesis
Can unsupervised computational methods effectively uncover trait heterogeneity in genotypic data?
Conclusion
Bayesian Classification is a promising method for identifying trait heterogeneity in genetic data, showing good control of false positive and negative rates.
Supporting Evidence
- Bayesian Classification achieved excellent recovery for 75% of datasets under the simplest genetic model.
- Fuzzy k-Modes Clustering performed best on the most complex genetic model.
- Bayesian Classification controlled the false positive rate at three percent or less across all significance levels.
Takeaway
This study looked at different ways to group genetic data to find hidden traits, and one method worked really well at finding these traits.
Methodology
The study compared three clustering methods: Bayesian Classification, Hypergraph-Based Clustering, and Fuzzy k-Modes Clustering using simulated genotypic data.
Potential Biases
Potential biases may arise from the choice of clustering method and the simulated data characteristics.
Limitations
The performance of clustering methods may vary with different genetic models and parameter settings.
Participant Demographics
The study used simulated datasets with varying sample sizes of affected individuals.
Statistical Information
P-Value
0.01
Confidence Interval
0.90
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website