Learning biophysically-motivated parameters for alpha helix prediction
2007

Predicting Protein Helices with a New Method

Sample size: 300 publication Evidence: moderate

Author Information

Author(s): Gassend Blaise, O'Donnell Charles W, Thies William, Lee Andrew, van Dijk Marten, Devadas Srinivas

Primary Institution: Massachusetts Institute of Technology

Hypothesis

Can a biophysically-motivated energy model improve the prediction of alpha helices in proteins?

Conclusion

The method shows promise for predicting protein secondary structure with competitive accuracy using only 302 parameters.

Supporting Evidence

  • The model achieved a Qα value of 77.6% and an SOVα value of 73.4%.
  • The method does not rely on external databases for predictions.
  • It uses only 302 parameters, making it simpler than many existing models.

Takeaway

This study created a new way to guess how proteins fold, focusing on a specific part called alpha helices, and it works pretty well with fewer rules.

Methodology

The study used a support vector machine to optimize a cost function based on biophysical principles for predicting alpha helices.

Limitations

The method is currently limited to predicting only alpha helices and may not generalize to other secondary structures.

Statistical Information

P-Value

p<0.05

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1186/1471-2105-8-S5-S3

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication