Predicting Protein Disorder Based on Length
Author Information
Author(s): Peng Kang, Predrag Radivojac, Slobodan Vucetic, A Keith Dunker, Zoran Obradovic
Primary Institution: Temple University
Hypothesis
The amino acid compositions and sequence properties of intrinsically disordered regions vary with length, affecting prediction accuracy.
Conclusion
The VSL2 predictors can accurately identify disordered regions of any length and improve upon previous models by addressing length-dependent prediction issues.
Supporting Evidence
- The VSL2 predictors achieved well-balanced prediction accuracies of 81% on both short and long disordered regions.
- VSL2 predictors were significantly more accurate than several existing predictors of intrinsic protein disorder.
- The success of the VSL2 predictors confirmed the differences in amino acid compositions between short and long disordered regions.
Takeaway
This study created new tools to better predict parts of proteins that don't have a fixed shape, helping scientists understand how these flexible parts work.
Methodology
The study developed two predictor models, VSL2-M1 and VSL2-M2, using specialized predictors for short and long disordered regions, integrated through a meta predictor.
Potential Biases
Potential mislabeling of disordered regions due to reliance on missing electron density in X-ray structures.
Limitations
The study did not include very short disordered regions (1-3 residues) in training or accuracy estimation, which may affect the generalizability of the findings.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website