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

Electronic van der Waals surface property descriptors and genetic algorithms for developing structure-activity

Barry K Lavine1, Charles E Davidson, Curt Breneman

  • 1Department of Chemistry, Clarkson University, Potsdam, New York 13699-5810.

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

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

Artificial Neural Networks for Discrimination of Automotive Clear Coats by Vehicle Manufacturer.

Sensors (Basel, Switzerland)·2026
Same author

Thermodynamic and Kinetic Characterization of Colloidal Polymers of <i>N</i>-Isopropylacrylamide and Alkyl Acrylic Acids for Optical pH Sensing.

Molecules (Basel, Switzerland)·2025
Same author

Alternating and Modified Alternating Least Squares Applied to Raman Spectra of Finished Gasolines.

Applied spectroscopy·2024
Same author

Opportunities for targeted therapies: trametinib as a therapeutic approach to canine oral squamous cell carcinomas.

Research square·2024
Same author

Supervised discretization for decluttering classification models.

The Analyst·2023
Same author

Raman Spectroscopy to Enhance Investigative Lead Information in Automotive Clearcoats.

Applied spectroscopy·2023

Machine learning accelerates the design of new odorants by developing a novel structure-activity correlation methodology. This approach uses advanced molecular descriptors and a genetic algorithm to streamline the discovery and development process.

Area of Science:

  • Computational chemistry
  • Machine learning
  • Drug discovery

Background:

  • Designing new odorants is traditionally a slow, expensive, and labor-intensive process.
  • Existing quantitative structure-activity relationship (QSAR) methods often struggle with structurally diverse datasets.

Purpose of the Study:

  • To develop an intelligent methodology for designing novel odorants with specific properties.
  • To streamline the odorant discovery and development pipeline using machine learning.

Main Methods:

  • Utilized large olfactory databases from scientific literature as input.
  • Employed an enhanced Transferable Atom Equivalent (TAE) methodology to generate molecular descriptors (PEST, WCD, TAE histogram).
  • Developed a genetic algorithm for pattern recognition to select optimal descriptors for class separation and data clustering.

Related Experiment Videos

Main Results:

  • The new methodology effectively characterizes molecules using shape-aware, electron density-based descriptors.
  • Overcame limitations of traditional fragment-based descriptors, especially for structurally varied molecules.
  • The genetic algorithm successfully identified descriptors that enhance data clustering and class separation.

Conclusions:

  • The developed machine learning methodology facilitates intelligent odorant design.
  • This approach offers a more efficient and accurate alternative to traditional odorant discovery methods.
  • The use of advanced molecular descriptors and pattern recognition significantly improves structure-activity correlation.