Improved Disorder Prediction by Combination of Orthogonal Approaches
2009

Improved Disorder Prediction

Sample size: 298 publication Evidence: high

Author Information

Author(s): Avner Schlessinger, Marco Punta, Guy Yachdav, Laszlo Kajan, Burkhard Rost

Primary Institution: Columbia University

Hypothesis

A combination of several orthogonal methods will capture many types of disorder at improved performance without sacrificing the distinction of the type of disorder that is detected.

Conclusion

The new method, MD, significantly outperformed each of its constituents and other commonly used prediction methods.

Supporting Evidence

  • MD outperformed its original methods in sustained cross-validation.
  • MD identified unique proteins that were missed by all other methods.
  • MD achieved an AUC of 0.80, indicating high accuracy.

Takeaway

Scientists created a new tool that helps predict which parts of proteins are disordered, and it works better than older methods.

Methodology

The study used a novel META-Disorder prediction method that combines various sources of information from different prediction methods.

Potential Biases

The method may over-predict short stretches and under-predict long regions due to limitations in the input data.

Limitations

Some aspects of disorder remain untapped, and the averaging of predictions may lose unique findings from original methods.

Statistical Information

P-Value

p<0.05

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1371/journal.pone.0004433

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication