Improving False Discovery Rate Estimates with Constrained Regression Recalibration
Author Information
Author(s): Li Juntao, Paramita Puteri, Choi Kwok Pui, Karuturi R Krishna Murthy
Primary Institution: Genome Institute of Singapore
Hypothesis
Can the empirical distribution of p-values be recalibrated to improve false discovery rate (FDR) estimates?
Conclusion
The ConReg-R method significantly improves FDR estimation on various gene expression datasets.
Supporting Evidence
- The ConReg-R method was shown to improve FDR estimation on simulated data.
- FDR estimation errors were significantly reduced after applying ConReg-R.
- ConReg-R was effective in recalibrating p-values from various gene expression datasets.
Takeaway
This study created a new method to make p-values more accurate, which helps scientists avoid mistakes when deciding if their results are real.
Methodology
The study used a new method called Constrained Regression Recalibration (ConReg-R) to model and recalibrate p-values from multiple hypothesis testing.
Potential Biases
Potential biases may arise from the assumptions made about the distribution of p-values.
Limitations
The method may not be valid for all datasets, especially if the p-values do not follow the expected distribution.
Statistical Information
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website