Query Large Scale Microarray Compendium Datasets Using a Model-Based Bayesian Approach with Variable Selection
2009

Querying Microarray Datasets with a Bayesian Approach

Sample size: 4217 publication 10 minutes Evidence: high

Author Information

Author(s): Hu Ming Qin, Zhaohui S.

Primary Institution: University of Michigan

Hypothesis

Statistically significant correlation can still be detected using microarray data, but strong correlation will be confined to a subset of samples/experimental conditions.

Conclusion

The proposed Bayesian model-based query algorithm effectively identifies co-expressed genes in large microarray datasets.

Supporting Evidence

  • BEST identified 28 target genes, 27 of which were in the RegulonDB target set.
  • BEST achieved an AUC of 0.87, significantly higher than other methods.
  • BEST is robust against added noise and complications in data.

Takeaway

This study created a smart tool that helps scientists find genes that work together by looking at their activity in different situations.

Methodology

A model-based query tool using Bayesian methods to identify genes with correlated expression profiles under specific experimental conditions.

Potential Biases

Potential biases due to measurement errors and the complexity of regulatory mechanisms.

Limitations

The model assumes all columns are independent and does not account for covariance, which may limit accuracy.

Participant Demographics

The study focused on Escherichia coli gene expression data.

Statistical Information

P-Value

1.27×10−6

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1371/journal.pone.0004495

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication