Using Hidden Markov Models to Analyze Genomic Tiling Microarrays
Author Information
Author(s): Kasper Munch, Paul P. Gardner, Peter Arctander, Anders Krogh
Primary Institution: University of Copenhagen
Hypothesis
Can a hidden Markov model improve the prediction of expression from genomic tiling microarray data?
Conclusion
The adaptive modeling approach used in ExpressHMM is superior to previous methods in predicting expressed and non-expressed regions.
Supporting Evidence
- The ExpressHMM method showed improved sensitivity and specificity compared to previous methods.
- ExpressHMM was tested on data from ten Human chromosomes.
- The model adapts to the data, avoiding arbitrary thresholds used in earlier approaches.
Takeaway
This study created a smart computer program that helps scientists figure out which parts of DNA are active by looking at special data from DNA tests.
Methodology
The study used a hidden Markov model trained on genomic tiling array data to categorize probes as expressed or non-expressed.
Potential Biases
The model may be biased due to reliance on incomplete or dubious expression annotations.
Limitations
The model's training relies on potentially inaccurate expression annotations, which may affect its predictions.
Participant Demographics
The study analyzed data from Human neuroblastoma cells (SK-N-AS cell line).
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website