A new pairwise kernel for biological network inference with support vector machines
2007

New Kernel for Biological Network Inference

publication Evidence: high

Author Information

Author(s): Jean-Philippe Vert, Jian Qiu, William S. Noble

Primary Institution: Centre for Computational Biology, Ecole des Mines de Paris; Department of Genome Sciences, University of Washington; Department of Computer Science and Engineering, University of Washington

Hypothesis

Can a new pairwise kernel improve the inference of biological networks using support vector machines?

Conclusion

The metric learning pairwise kernel provides state-of-the-art results for inferring biological networks from genomic data.

Supporting Evidence

  • The new kernel improves upon the state-of-the-art SVM for indirect inference.
  • Combining the new kernel with existing methods leads to better results.
  • The method was validated on real biological networks and genomic datasets.

Takeaway

This study created a new way to predict how proteins interact by using a special math tool called a kernel, which helps scientists understand biological networks better.

Methodology

The study used support vector machines with a new pairwise kernel to infer relationships in biological networks from various genomic datasets.

Limitations

The proposed method may not be applicable when the training data does not represent a complete set of high-confidence edges.

Digital Object Identifier (DOI)

10.1186/1471-2105-8-S10-S8

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication