Probabilistic prediction and ranking of human protein-protein interactions
2007

Predicting Human Protein-Protein Interactions

Sample size: 37606 publication 10 minutes Evidence: moderate

Author Information

Author(s): Scott Michelle S, Barton Geoffrey J

Primary Institution: University of Dundee, Scotland, UK

Hypothesis

Can we improve the prediction of human protein-protein interactions using a probabilistic framework that combines various protein features?

Conclusion

The study successfully predicts over 37,000 human protein interactions, significantly increasing the coverage of the human interaction map.

Supporting Evidence

  • The method identified 37,606 human interactions, with 32,892 being novel.
  • The false positive rate of the new method is estimated to be below 80%.
  • Independent validation of a subset of predicted interactions was performed.

Takeaway

The researchers created a method to guess how proteins in humans interact with each other, finding many new connections that weren't known before.

Methodology

The study used a Bayesian framework to integrate various protein features, including expression data, orthology, and local network topology, to predict interactions.

Potential Biases

Potential biases may arise from the reliance on existing datasets and the assumptions made in the probabilistic model.

Limitations

The predictions may still have a high false positive rate and the method relies on the quality of the input data.

Statistical Information

P-Value

p<0.05

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1186/1471-2105-8-239

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication