Predictive Power Estimation Algorithm (PPEA) - A New Algorithm to Reduce Overfitting for Genomic Biomarker Discovery
2011

New Algorithm for Genomic Biomarker Discovery

Sample size: 40 publication 10 minutes Evidence: high

Author Information

Author(s): Liu Jiangang, Jolly Robert A., Smith Aaron T., Searfoss George H., Goldstein Keith M., Uversky Vladimir N., Dunker Keith, Li Shuyu, Thomas Craig E., Wei Tao

Primary Institution: Lilly Research Laboratories

Hypothesis

Can the Predictive Power Estimation Algorithm (PPEA) reduce overfitting in genomic biomarker discovery?

Conclusion

The PPEA model effectively addresses the overfitting problem and facilitates genomic biomarker discovery for predictive toxicology and drug responses.

Supporting Evidence

  • PPEA can quickly derive a reliable rank order of predictive power of individual transcripts.
  • The top ranked transcripts tend to be functionally related to the phenotype they are intended to predict.
  • Using only the most predictive top ranked transcripts greatly facilitates development of multiplex assays.
  • A small number of genes identified from the top-ranked transcripts are highly predictive of phenotype.

Takeaway

Scientists created a new computer program to help find important genes that can predict if a medicine will harm the liver, making it easier to test new drugs safely.

Methodology

The study used a novel algorithm called PPEA, which applies two-way bootstrapping to evaluate and rank the predictive power of individual genes.

Potential Biases

Different runs of the algorithm may select different features if the number of iterations is small.

Limitations

The algorithm may not find all interesting genes due to its heuristic and suboptimal search method.

Participant Demographics

Rats were used in the study, with a focus on liver toxicity.

Statistical Information

P-Value

0.01

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1371/journal.pone.0024233

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication