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

Evaluation of machine-learning methods for ligand-based virtual screening.

Beining Chen1, Robert F Harrison, George Papadatos

  • 1Department of Chemistry, University of Sheffield, Western Bank, Sheffield, UK.

Journal of Computer-Aided Molecular Design
|January 6, 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

Synthesis of Phenyl 2-Acetamidoselenogalactoside Mimetics and Interaction with Amyloid β<sub>1-42</sub>.

Pharmaceuticals (Basel, Switzerland)·2026
Same author

Citric acid-enhanced chitosan/hyaluronic acid interpenetrating hydrogel network as a bioinspired adhesive for osteochondral scaffold fabrication.

International journal of biological macromolecules·2026
Same author

Parameter Optimisation in 3D Extrusion Printing of Polyhydroxybutyrate Using Design of Experiment Methodology.

Journal of functional biomaterials·2026
Same author

A larger defect size and more prior surgical procedures are negatively associated with radiological outcome 10 years after tibiofemoral matrix-induced autologous chondrocyte implantation.

Knee surgery, sports traumatology, arthroscopy : official journal of the ESSKA·2025
Same author

Discovery and Characterization of Zilurgisertib, a Potent and Selective Inhibitor of Activin Receptor-like Kinase‑2 (ALK2) for the Treatment of Fibrodysplasia Ossificans Progressiva.

ACS medicinal chemistry letters·2025
Same author

10-Year Clinical and MRI-Based Outcomes of a Randomized Controlled Trial Evaluating a 6-Week Return to Full Weightbearing After Matrix-Induced Autologous Chondrocyte Implantation.

Orthopaedic journal of sports medicine·2025
Same journal

Topological data analysis for antibody-drug conjugate payload discovery: a computational framework for mechanistic classification and target validation.

Journal of computer-aided molecular design·2026
Same journal

Commentary on the fundamentals and development of artificial intelligence models in the life sciences and best research practices.

Journal of computer-aided molecular design·2026
Same journal

RANQSAR: a standalone open-source application for reproducible machine learning-based QSAR analysis.

Journal of computer-aided molecular design·2026
Same journal

Integrating evolutionary and compositional features with ML and DL for robust and interpretable druggable protein prediction.

Journal of computer-aided molecular design·2026
Same journal

QUAD: a composite risk framework integrating uncertainty, applicability domain, and model disagreement for reliable QSAR predictions.

Journal of computer-aided molecular design·2026
Same journal

Comparative quantum-chemical investigation of 2-chloro-N-(4-methoxyphenyl)acetamide and 2-(4-methoxyphenylamino)-2-oxoethyl meth/acrylate: DFT, TD-DFT, and non-covalent interaction analyses.

Journal of computer-aided molecular design·2026
See all related articles

Machine learning aids virtual screening by analyzing molecular structures. Kernel methods and naive Bayesian classifiers (NBC) show comparable performance, though group fusion excels with limited active data in diverse datasets.

Area of Science:

  • Computational chemistry
  • Cheminformatics
  • Machine learning

Background:

  • Virtual screening is crucial for drug discovery, leveraging machine learning to analyze molecular structures.
  • Kernel methods and naive Bayesian classifiers (NBC) are established techniques for structure-activity relationship analysis.

Purpose of the Study:

  • To evaluate kernel discrimination and NBC methods for virtual screening.
  • To compare the performance of a novel kernel method with existing ones.
  • To assess NBC performance with limited active training data.

Main Methods:

  • Utilized kernel discrimination for processing molecules with diverse descriptors (binary, integer, real-valued).
  • Implemented a naive Bayesian classifier (NBC) for virtual screening tasks.

Related Experiment Videos

  • Compared performance against a kernel method specialized for binary fingerprints.
  • Investigated group fusion as an alternative for imbalanced datasets.

    • Main Results:

    • The novel kernel method demonstrated comparable screening performance to existing fingerprint-specific kernels.
    • NBC performance was evaluated under conditions of scarce active molecules in training sets.
    • Group fusion exhibited superior screening performance for structurally heterogeneous datasets with few active compounds.

    Conclusions:

    • Kernel methods offer flexibility in handling various molecular descriptor types for virtual screening.
    • Naive Bayesian classifiers may be outperformed by simpler methods like group fusion when training data is limited and diverse.