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

Spline-fitting with a genetic algorithm: a method for developing classification structure-activity relationships.

Jeffrey J Sutherland1, Lee A O'Brien, Donald F Weaver

  • 1Departments of Chemistry and Pathology, Queen's University, Kingston, Ontario, Canada K7L 3N6.

Journal of Chemical Information and Computer Sciences
|November 25, 2003
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

Computational Insights into the Antiviral Properties of the Antimicrobial Peptide β-Amyloid.

The journal of physical chemistry. B·2026
Same author

β-Alanine Is an Unexploited Neurotransmitter in the Pathogenesis and Treatment of Alzheimer's Disease.

NeuroSci·2026
Same author

The Cause-and-Consequence Relationships between Domestic Abuse and Alzheimer's Disease: Identification of Five Subtypes.

Current Alzheimer research·2026
Same author

Dementia Care Research and Psychosocial Factors.

Alzheimer's & dementia : the journal of the Alzheimer's Association·2025
Same author

Basic Science and Pathogenesis.

Alzheimer's & dementia : the journal of the Alzheimer's Association·2025
Same author

Marketing Science: A Missing Piece in the Fight Against Alzheimer's.

The Canadian journal of neurological sciences. Le journal canadien des sciences neurologiques·2025

A new classification method, SFGA, shows improved predictive power for structure-activity relationship models compared to recursive partitioning (RP) and SIMCA. A consensus approach combining all methods yielded the best overall results for compound classification.

Area of Science:

  • Computational chemistry
  • Cheminformatics
  • Drug discovery

Background:

  • Structure-activity relationship (SAR) models are crucial for drug discovery.
  • Categorical target properties require specialized classification methods for SAR modeling.
  • Existing methods like recursive partitioning (RP) and SIMCA have limitations.

Purpose of the Study:

  • To introduce and evaluate a novel classification method, SFGA (descriptor splines fitted to activities with descriptors selected using a genetic algorithm).
  • To compare the predictive performance and stability of SFGA against RP and SIMCA.
  • To assess the utility of a consensus approach combining multiple classification methods.

Main Methods:

  • SFGA method development: fitting descriptor splines to activities, with descriptor selection via genetic algorithm.

Related Experiment Videos

  • Application to five compound series: COX-2 inhibitors, BZR ligands, ER ligands, DHFR inhibitors, and MAO inhibitors.
  • Validation using distinct test sets (cherry-picked and random) and stability analysis.

    • Main Results:

    • SFGA demonstrated superior predictive performance across most datasets compared to RP and SIMCA.
    • SIMCA provided the most predictive models for the DHFR inhibitor set.
    • RP generally yielded the least predictive models.
    • SIMCA models exhibited the highest stability, followed by SFGA and RP.
    • A consensus approach combining SFGA, RP, and SIMCA consistently outperformed individual methods.

    Conclusions:

    • SFGA offers a valuable alternative for developing predictive SAR models with categorical endpoints.
    • Model stability is a critical consideration, with SIMCA showing the highest robustness.
    • Ensemble or consensus modeling strategies can enhance predictive accuracy and reliability in cheminformatics.