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

Contact prediction using mutual information and neural nets.

George Shackelford1, Kevin Karplus

  • 1Department of Biomolecular Engineering, University of California, Santa Cruz, California 95064, USA.

Proteins
|October 13, 2007
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

Altering under-represented DNA sequences elevates bacterial transformation efficiency.

mBio·2023
Same author

Fallacy of the Unique Genome: Sequence Diversity within Single <i>Helicobacter pylori</i> Strains.

mBio·2017
Same author

Analysis of nanopore data using hidden Markov models.

Bioinformatics (Oxford, England)·2015
Same author

Error rates for nanopore discrimination among cytosine, methylcytosine, and hydroxymethylcytosine along individual DNA strands.

Proceedings of the National Academy of Sciences of the United States of America·2013
Same author

Complete genome sequence of Pyrobaculum oguniense.

Standards in genomic sciences·2013
Same author

Automated forward and reverse ratcheting of DNA in a nanopore at 5-Å precision.

Nature biotechnology·2012

Predicting protein structures is challenging without templates. This study introduces a neural network for residue-residue contact prediction, improving accuracy using correlated mutation statistics and amino acid propensities.

Area of Science:

  • Computational Biology
  • Structural Bioinformatics
  • Machine Learning in Biology

Background:

  • Protein structure prediction remains a significant challenge, especially for proteins lacking homologous templates.
  • Local structure predictions (secondary structure, residue burial) are reliable but insufficient for complete 3D models.
  • Residue-residue contact prediction offers a potential pathway to assemble local structures into tertiary models.

Purpose of the Study:

  • To develop an improved method for predicting residue-residue contacts in proteins.
  • To leverage correlated mutation detection and amino acid propensities within a neural network framework.
  • To evaluate the performance of the developed contact prediction method against existing approaches.

Main Methods:

  • A neural network was designed to predict contact probabilities between pairs of residue positions.

Related Experiment Videos

  • A novel statistic, the statistical significance of mutual information from multiple sequence alignments, was incorporated as a neural network input.
  • Additional features included amino acid distributions and local structure predictions.

    • Main Results:

    • The developed neural network effectively predicts residue-residue contacts.
    • The method demonstrated superior performance compared to other contact prediction methods evaluated at CASP7.
    • The approach, incorporating correlated mutation statistics and amino acid propensities, proved to be a strong contact predictor.

    Conclusions:

    • The developed neural network offers a promising advancement in residue-residue contact prediction.
    • This method shows potential for guiding the assembly of local structure predictions into complete protein tertiary structures.
    • Further research is needed to determine if contact predictions can enhance tertiary models in free-modeling domains.