Impact of False Positives and Negatives on Binding Target Analysis
Author Information
Author(s): Debayan Datta, Hongyu Zhao
Primary Institution: Yale University
Hypothesis
How do false positive and false negative rates affect the inference of binding target conservation across different conditions and species from ChIP-chip data?
Conclusion
The study shows that moderate false positive and false negative rates do not significantly alter the inference of conservation when there is a strong association among binding targets.
Supporting Evidence
- The study proposes a statistical method to account for false positives and negatives in ChIP-chip data analysis.
- Results indicate that high odds ratios suggest strong associations that are robust to moderate errors in data.
- The EM algorithm effectively infers true binding states from observed data despite the presence of noise.
Takeaway
This study looks at how mistakes in data can change our understanding of how genes are regulated. It finds that if the connections between genes are strong, small mistakes won't change our conclusions much.
Methodology
The study uses an Expectation Maximization approach to analyze ChIP-chip data and infer true binding states from observed counts.
Potential Biases
Potential underestimation of association due to non-independence of data points in closely associated experiments.
Limitations
The EM algorithm cannot estimate false positive and false negative rates due to limited degrees of freedom in the data.
Statistical Information
P-Value
0.001
Statistical Significance
p<0.001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website