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

Protein contact prediction using patterns of correlation.

Nicholas Hamilton1, Kevin Burrage, Mark A Ragan

  • 1Advanced Computational Modelling Centre, Department of Mathematics, The University of Queensland, St. Lucia, Queensland, Australia. nick@maths.uq.edu.au

Proteins
|July 29, 2004
PubMed
Summary

This study introduces a novel neural network method to predict protein residue contacts. By analyzing correlated mutations within sequence windows, the approach significantly enhances prediction accuracy for protein structures.

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

A mathematical perspective on hypothesis-driven model construction: A case study in pea.

Mathematical biosciences·2026
Same author

"It Felt Good to Be Able to Say That Out Loud"-Therapeutic Alliance and Processes in AVATAR Therapy for People Who Hear Distressing Voices: Peer-Led Qualitative Study.

JMIR mental health·2026
Same author

Diagnosis and Management of Isolated Laryngeal Sarcoidosis: A Systematic Review.

Otolaryngology--head and neck surgery : official journal of American Academy of Otolaryngology-Head and Neck Surgery·2026
Same author

Author Correction: Evidence for improved DNA repair in the long-lived bowhead whale.

Nature·2025
Same author

Evidence for improved DNA repair in the long-lived bowhead whale.

Nature·2025
Same author

Persistent Hematuria With Late Diagnosis of Cardiac Origin: A Case Report.

Cureus·2025

Area of Science:

  • Computational biology
  • Structural bioinformatics
  • Machine learning in genomics

Background:

  • Predicting residue-residue contacts is crucial for protein structure determination.
  • Correlated mutation analysis is a common, yet often weak, predictor of residue contacts.

Purpose of the Study:

  • To develop and validate a new neural network-based method for predicting residue contact pairs in proteins.
  • To improve the accuracy of contact prediction by leveraging correlated mutation data within sequence windows.

Main Methods:

  • Utilizing a neural network trained on correlated mutation measures from windows of 5 residues.
  • Training dataset comprised 100 proteins; testing dataset included 1033 proteins of known structure.
  • Evaluating predictor performance on CASP5 proteins and analyzing accuracy across sequence lengths and secondary structures.

Related Experiment Videos

Main Results:

  • Achieved an average predictive accuracy of 21.7% for the top L/2 predictions and 30.7% for the top L/10 predictions.
  • Demonstrated improved accuracy when using multiple sequence alignments with numerous sequences for correlation calculations.
  • Observed consistent accuracy across varying sequence lengths but significant variation based on secondary structure.

Conclusions:

  • The developed neural network method offers a significant improvement over traditional correlated mutation analysis for predicting residue contacts.
  • The method's accuracy is influenced by sequence length and secondary structure, highlighting areas for future refinement.
  • Employing larger multiple sequence alignments enhances the predictive power of correlated mutation analysis.