Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

A composite score for predicting errors in protein structure models.

David Eramian1, Min-yi Shen, Damien Devos

  • 1Graduate Group in Biophysics, Department of Biopharmaceutical Sciences, University of California at San Francisco 94158, USA.

Protein Science : a Publication of the Protein Society
|June 6, 2006
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Extending structural surfaceomics to identify aberrant conformations of tumor surface proteins as potential immunotherapy targets.

bioRxiv : the preprint server for biology·2026
Same author

Systematic discovery of pro- and anti-HIV host factors in primary human CD4+ T cells.

Cell·2026
Same author

The scientific legacy of Martin Karplus from the perspective of his collaborators.

Biophysical journal·2026
Same author

The filamentous ultrastructure of the PopZ condensate is required for its cellular function.

Nature structural & molecular biology·2026
Same author

ModelCIF Update: Supporting Emerging Classes of Computational Macromolecular Models.

Journal of molecular biology·2026
Same author

IHMValidation: Assessment of Integrative Structure Models Deposited to the Protein Data Bank.

Journal of molecular biology·2025
Same journal

Macromolecular crowding inhibits degradation of alpha-synuclein amyloid fibrils induced by cathepsins and MMP9.

Protein science : a publication of the Protein Society·2026
Same journal

Sequence-encoded differences in the conformational ensembles of CITED transcriptional activation domains impact coactivator binding.

Protein science : a publication of the Protein Society·2026
Same journal

The phospholipid biosynthesis enzyme PlsB contains three distinct domains for membrane association, lysophosphatidic acid synthesis, and dimerization.

Protein science : a publication of the Protein Society·2026
Same journal

Structural basis of ligand selectivity in FAD/NAD(P)H-dependent dehydrogenases: insights from trypanothione reductase and type II NADH dehydrogenase.

Protein science : a publication of the Protein Society·2026
Same journal

Achieving protease substrate-specific inhibition by mAb dual functional selections.

Protein science : a publication of the Protein Society·2026
Same journal

How important are quantum mechanical effects in controlling biological functions: Enzymes, electron transfer and bird navigation.

Protein science : a publication of the Protein Society·2026
See all related articles

Accurately predicting protein model accuracy is crucial. A new composite scoring function, using support vector machine (SVM) regression, significantly improved identifying the most native-like protein models.

Area of Science:

  • Computational Biology
  • Structural Bioinformatics
  • Biophysics

Background:

  • Reliable prediction of protein model accuracy remains a challenge in structural biology.
  • Existing assessment scores often struggle to accurately rank models, especially in difficult comparative modeling scenarios.

Purpose of the Study:

  • To develop and evaluate a superior method for predicting protein model accuracy.
  • To identify the most native-like protein models from large sets of generated structures.

Main Methods:

  • Evaluated 24 individual scoring functions (physics-based, statistical, machine learning).
  • Constructed ~85,000 composite scoring functions using support vector machine (SVM) regression.
  • Tested scoring functions on 6000 comparative models for 20 protein targets with low sequence identity templates (<30%).

Related Experiment Videos

Main Results:

  • The best SVM-based composite score outperformed all individual scores.
  • Reduced the average difference (DeltaRMSD) between the best-ranked and lowest RMSD models from 0.63 Å to 0.45 Å.
  • Achieved a higher Pearson correlation coefficient (r=0.87) to RMSD compared to individual scores.

Conclusions:

  • A novel SVM-based composite scoring function significantly enhances the accuracy of protein model selection.
  • The best performing score combines multiple potentials (DOPE, MODPIPE, PSIPRED/DSSP) and is implemented in the SVMod program.
  • SVMod can be applied to various protein modeling tasks, including fold assignment, alignment, and loop modeling.