A constrained polynomial regression procedure for estimating the local False Discovery Rate
2007

New Method for Estimating Local False Discovery Rate

Sample size: 5000 publication Evidence: high

Author Information

Author(s): Cyril Dalmasso, Avner Bar-Hen, Philippe Broët

Primary Institution: Univ. Paris-Sud

Hypothesis

Can a new polynomial regression method improve the estimation of the local False Discovery Rate (lFDR) in genomic studies?

Conclusion

A novel and efficient procedure for estimating lFDR was developed and evaluated.

Supporting Evidence

  • The new estimator showed better performance in simulations compared to four existing methods.
  • The method was applied to real datasets from genomic studies, demonstrating its practical utility.
  • The polynomial regression approach allows for direct estimation of the local False Discovery Rate.

Takeaway

This study created a new way to figure out how likely it is that a gene is really important when looking at many genes at once.

Methodology

The study used a polynomial regression approach to estimate the local False Discovery Rate without distributional assumptions.

Potential Biases

The method may exhibit conservative bias in estimating the lFDR.

Limitations

The method assumes uniform distribution of p-values under the null hypothesis and may not perform well if this assumption is violated.

Digital Object Identifier (DOI)

10.1186/1471-2105-8-229

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication