Comparison of normalization methods for CodeLink Bioarray data
2005

Comparing Methods to Normalize CodeLink Bioarray Data

Sample size: 36 publication Evidence: high

Author Information

Author(s): Wu Wei, Dave Nilesh, Tseng George C, Richards Thomas, Xing Eric P, Kaminski Naftali

Primary Institution: University of Pittsburgh Medical Center

Hypothesis

Which normalization method is most effective for CodeLink Bioarray data?

Conclusion

CyclicLoess is the most effective normalization method for CodeLink Bioarray data, significantly improving variability reduction and signal retention compared to Median.

Supporting Evidence

  • CyclicLoess consistently reduced variability more effectively than Median.
  • CyclicLoess revealed the largest number of differentially expressed genes in both datasets.
  • Quantile normalization did not consistently retain signal in the time course dataset.

Takeaway

This study looked at different ways to clean up data from a specific type of gene test, and found that one method worked best to make the results clearer.

Methodology

Five normalization methods (Median, CyclicLoess, Quantile, Iset, and Qspline) were compared using two real datasets from CodeLink Bioarrays.

Potential Biases

The performance of Iset varied depending on the choice of baseline array.

Limitations

The sample sizes in the time course dataset were unbalanced, and no spike-in datasets were available for validation.

Participant Demographics

The time course dataset included control and treatment groups of rats, while the IPF dataset included control patients and patients with idiopathic pulmonary fibrosis.

Statistical Information

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1186/1471-2105-6-309

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication