Comparison of Statistical Data Models for Identifying Differentially Expressed Genes Using a Generalized Likelihood Ratio Test
2008

Comparing Statistical Models for Gene Expression Analysis

Sample size: 2000 publication Evidence: high

Author Information

Author(s): Seng Kok-Yong, Glenny Robb W., Madtes David K., Spilker Mary E., Vicini Paolo, Gharib Sina A.

Primary Institution: University of Washington

Hypothesis

How do different statistical error structures affect the performance of the generalized likelihood ratio test in identifying differentially expressed genes?

Conclusion

The generalized likelihood ratio tests outperform the traditional t-test for detecting differential gene expression, especially when considering the underlying error structure.

Supporting Evidence

  • The GLR tests showed higher power in detecting differential gene expression compared to the t-test.
  • The identity of the underlying error structure significantly influenced the performance of the GLR tests.
  • Signal-to-noise ratio was found to be more critical than sample replication in identifying statistically significant differential gene expression.

Takeaway

This study shows that using better math models can help scientists find important gene changes more accurately, especially when they have less data.

Methodology

The study compared different statistical error structures using simulated microarray data to evaluate their performance in identifying differentially expressed genes.

Limitations

The study's findings may not apply to all types of microarray platforms and did not address multiple comparison issues.

Participant Demographics

Four male Balb/C mice were used in the experiments.

Statistical Information

Statistical Significance

p<0.05

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication