Methods to Improve ChIP-Seq Data Analysis
Author Information
Author(s): David A. Nix, Samir J. Courdy, Kenneth M. Boucher
Primary Institution: Huntsman Cancer Institute, University of Utah
Hypothesis
Can algorithms and software be developed to reduce false positives and estimate confidence in ChIP-Seq peaks?
Conclusion
The developed methods show promise for reducing false positives and estimating confidence in ChIP-Seq data without prior knowledge of the target.
Supporting Evidence
- The use of control input data more than doubled the recovery of ChIP-Seq peaks at a 5% false discovery rate.
- Both a binomial p-value/q-value and an empirical FDR were found to be more reliable estimators of confidence than a global Poisson p-value.
- The methods were validated using simulated spike-in datasets.
Takeaway
This study created new tools to help scientists find important DNA binding sites more accurately, making it easier to understand how genes are controlled.
Methodology
The study developed algorithms and software to analyze ChIP-Seq data, comparing methods using simulated datasets.
Potential Biases
Systematic bias in ChIP-Seq data can lead to false positives if not controlled.
Limitations
The methods may not perform well with datasets lacking control input data.
Statistical Information
P-Value
0.001
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website