Dimension reduction with gene expression data using targeted variable importance measurement
2011

Improving Gene Selection with TMLE-VIM

Sample size: 133 publication 10 minutes Evidence: moderate

Author Information

Author(s): Wang Hui, van der Laan Mark J

Primary Institution: Stanford University and University of California Berkeley

Hypothesis

Can the TMLE-VIM method improve the selection of genes for predicting clinical outcomes in breast cancer?

Conclusion

The TMLE-VIM method enhances the quality of gene selection for predicting clinical responses in breast cancer patients.

Supporting Evidence

  • TMLE-VIM identified more genes than traditional methods.
  • The TMLE-VIM method showed better prediction accuracy in simulations.
  • The method effectively reduced the number of falsely associated variables.

Takeaway

This study shows a new way to pick important genes that help doctors understand how well breast cancer patients will respond to treatment.

Methodology

The study used simulations and data analyses to compare the TMLE-VIM method with traditional methods for gene selection.

Potential Biases

Potential bias due to the choice of initial estimators and confounding variables.

Limitations

The TMLE-VIM method may be sensitive to overfitting and requires careful selection of parameters.

Participant Demographics

133 breast cancer patients with gene expression data.

Statistical Information

P-Value

0.005

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1186/1471-2105-12-312

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication