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-Ligand Scoring with Convolutional Neural Networks.

Matthew Ragoza, Joshua Hochuli, Elisa Idrobo1

  • 1Department of Computer Science, The College of New Jersey , Ewing, New Jersey 08628, United States.

Journal of Chemical Information and Modeling
|April 4, 2017
PubMed
Summary

Convolutional neural network (CNN) scoring functions improve computational drug discovery by accurately predicting protein-ligand binding. These deep learning models outperform traditional methods in ranking binding poses for virtual screening.

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

FlowMol3: flow matching for 3D <i>de novo</i> small-molecule generation.

Digital discovery·2026
Same author

A Cationic Amphiphilic Drug (CAD) Defense System in the Nematode <i>Caenorhabditis elegans</i>.

bioRxiv : the preprint server for biology·2026
Same author

GEOM-drugs revisited: toward more chemically accurate benchmarks for 3D molecule generation.

Digital discovery·2025
Same author

Sliding Window Interaction Grammar (SWING): a generalized interaction language model for peptide and protein interactions.

Nature methods·2025
Same author

CACHE Challenge #2: Targeting the RNA Site of the SARS-CoV-2 Helicase Nsp13.

Journal of chemical information and modeling·2025
Same author

AI-driven discovery of synergistic drug combinations against pancreatic cancer.

Nature communications·2025

Area of Science:

  • Computational chemistry
  • Structural biology
  • Bioinformatics

Background:

  • Structure-based drug design utilizes scoring functions to predict protein-ligand binding affinities and poses.
  • Deep machine learning leverages vast protein-ligand binding data for enhanced scoring.
  • Traditional scoring functions like AutoDock Vina have limitations in accuracy.

Purpose of the Study:

  • To develop and evaluate convolutional neural network (CNN) scoring functions for protein-ligand interactions.
  • To assess the performance of CNNs in discriminating correct from incorrect binding poses.
  • To compare CNN scoring function efficacy against established methods like AutoDock Vina.

Main Methods:

  • Inputting comprehensive three-dimensional (3D) representations of protein-ligand interactions into CNN models.

Related Experiment Videos

  • Training and optimizing CNNs to identify key interaction features correlating with binding affinity.
  • Utilizing datasets of known binders and non-binders for model discrimination.
  • Evaluating CNN performance in pose prediction and virtual screening tasks.

    • Main Results:

    • CNN scoring functions automatically learn critical features of protein-ligand interactions.
    • The developed CNN models demonstrate superior performance in ranking binding poses compared to AutoDock Vina.
    • CNNs effectively discriminate between correct and incorrect binding poses and identify known binders.

    Conclusions:

    • CNN-based scoring functions represent a significant advancement in computational drug discovery.
    • Deep learning approaches offer improved accuracy and efficiency for structure-based drug design.
    • These findings suggest a promising future for AI-driven methods in identifying novel drug candidates.