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 dissimilarity matrix between protein atom classes based on Gaussian mixtures.

Ville-Veikko Rantanen1, Mats Gyllenberg, Timo Koski

  • 1Department of Mathematics, University of Turku, FIN-20014 Turku, Finland.

Bioinformatics (Oxford, England)
|September 10, 2002
PubMed
Summary

This study refines protein atom classification for better protein-ligand interaction predictions. By analyzing dissimilarities between atom classes, we propose a reduced classification system to enhance predictive accuracy.

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

Uncovering Gaps in Obesity Medicine Competencies: Insights from Ten U.S. Medical Schools.

Medical science educator·2026
Same author

Spirituality as a Component of Student Resilience and Well-Being in Standards 2025.

American journal of pharmaceutical education·2026
Same author

Metabolism matters: Interspecies variability and ecological traits of biotransformation kinetics.

Ecotoxicology and environmental safety·2026
Same author

The Role of the Omega Subzone in Determining the Membership of a Protein in One of the Two Families of the LexA/Signal Peptidase-like Superfamily.

International journal of molecular sciences·2026
Same author

Novel, Stable, and Well-Secreted Single-Chain Fc Forms of Programmed Death-1 Immune Checkpoint Inhibitor Efficiently Block PD-1/PD-L1 Interaction.

Human gene therapy·2026
Same author

Recurrent cancer-associated ERBB4 mutations are transforming and confer resistance to targeted therapies.

Molecular oncology·2025

Area of Science:

  • Computational Biology
  • Structural Biology
  • Bioinformatics

Background:

  • Previous work created a protein atom-ligand interaction library using 3D structural data.
  • Gaussian mixture models and expectation-maximization were used to model protein atom locations.
  • Validation suggested the existing 24 protein atom classes were too numerous for optimal prediction.

Purpose of the Study:

  • To derive a dissimilarity matrix for comparing and merging 24 protein atom classes.
  • To guide the readjustment of protein atom classification.
  • To improve the accuracy of protein-ligand interaction predictions.

Main Methods:

  • Derived a dissimilarity matrix using Jeffreys' distances between Gaussian mixture models.
  • Applied hierarchical clustering to analyze the dissimilarity data.

Related Experiment Videos

  • Utilized multidimensional scaling analysis to further investigate class relationships.

    • Main Results:

    • Quantified dissimilarities between 24 pre-defined protein atom classes.
    • Identified relationships between different protein atom classes.
    • Provided data to support a reduced and optimized protein atom classification system.

    Conclusions:

    • The derived dissimilarity matrix and subsequent analyses offer insights into protein atom class relationships.
    • Readjusting the protein atom classification based on these findings can improve predictive models.
    • This work contributes to more accurate computational predictions of protein-ligand interactions.