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

Multi-class protein fold recognition using support vector machines and neural networks.

C H Ding1, I Dubchak

  • 1NERSC Division, Lawrence Berkeley National Laboratory, University of California, Berkeley, CA 94720, USA.

Bioinformatics (Oxford, England)
|April 13, 2001
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

Fungal coinfections among patients with COVID-19: demographics, risk factors and outcomes.

The Malaysian journal of pathology·2026
Same author

Antifungal susceptibility of non-albicans Candida species from blood samples.

The Malaysian journal of pathology·2025
Same author

Acute rotavirus gastroenteritis in children less than 5 years old: Salient clinical and laboratory features.

The Malaysian journal of pathology·2025
Same author

Bacteremia due to the fastidious bacterium Granulicatella adiacens: A diagnosis that was almost missed.

Tropical biomedicine·2025
Same author

Marrow talaromycosis as the initial presentation in a case of Burkitt lymphoma.

The Malaysian journal of pathology·2024
Same author

Septicaemia due to Vibrio vulnificus: A tropical infection not to be taken lightly.

Tropical biomedicine·2024

New protein fold recognition methods significantly improve accuracy by addressing false positives. Advanced classification techniques like Support Vector Machines (SVM) and Neural Networks (NN) enhance protein structure discovery without sequence similarity.

Area of Science:

  • Computational biology
  • Structural bioinformatics

Background:

  • Protein fold recognition is crucial for structure discovery, independent of sequence similarity.
  • Existing methods often suffer from 'False Positives' due to the one-against-others approach.

Purpose of the Study:

  • To investigate novel multi-class classification methods for practical protein fold recognition.
  • To enhance prediction accuracy and address limitations of current discriminative models.

Main Methods:

  • Exploration of unique one-against-others and all-against-all classification strategies.
  • Utilized Support Vector Machine (SVM) and Neural Network (NN) as base classifiers.
  • Implemented majority voting with combined scores from multiple parameter datasets.

Main Results:

Related Experiment Videos

  • Achieved 14-110% improvement in prediction accuracy on a dataset of 27 SCOP folds.
  • SVM demonstrated fast convergence and high accuracy.
  • Overall system reached 56% fold prediction accuracy on a challenging test dataset with low sequence identity.

Conclusions:

  • Novel classification methods and ensemble techniques effectively improve protein fold recognition accuracy.
  • The developed system offers a robust approach for structure discovery with limited sequence homology.