Parameter estimation for robust HMM analysis of ChIP-chip data
2008

Improving Hidden Markov Model Analysis of ChIP-chip Data

publication Evidence: high

Author Information

Author(s): Peter Humburg, David Bulger, Glenn Stone

Primary Institution: Department of Statistics, Macquarie University

Hypothesis

Can a hidden Markov model with robust parameter estimation improve the analysis of ChIP-chip data?

Conclusion

The developed hidden Markov model significantly outperforms established methods like TileMap in analyzing histone modification studies.

Supporting Evidence

  • The new model showed improved sensitivity and specificity compared to TileMap.
  • The model was able to identify enriched regions with high accuracy.
  • Parameter estimation from data led to better performance than ad hoc estimates.

Takeaway

This study created a new way to analyze DNA data that helps scientists find important areas more accurately, especially when looking at how proteins interact with DNA.

Methodology

The study used a hidden Markov model with t emission distributions and maximum likelihood estimation for parameter estimation.

Potential Biases

The reliance on simulated data may introduce bias in performance assessment.

Limitations

The model may produce some short regions that appear to be false positives, which could be artifacts from the simulation process.

Statistical Information

P-Value

2 × 10^-15

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1186/1471-2105-9-343

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication