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

Sample-distance partial least squares: PLS optimized for many variables, with application to CoMFA

B L Bush1, R B Nachbar

  • 1Merck Research Laboratories, Merck & Co., Inc., Rahway, NJ 07065.

Journal of Computer-Aided Molecular Design
|October 1, 1993
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

Comparison of knowledge-based and distance geometry approaches for generation of molecular conformations.

Journal of chemical information and computer sciences·2001
Same author

Latent semantic structure indexing (LaSSI) for defining chemical similarity.

Journal of medicinal chemistry·2001
Same author

Chemical similarity searches using latent semantic structural indexing (LaSSI) and comparison to TOPOSIM.

Journal of medicinal chemistry·2001
Same author

The statistical-thermodynamic basis for computation of binding affinities: a critical review.

Biophysical journal·1997
Same author

Two new criteria for choosing sample size in combinatorial chemistry.

Journal of medicinal chemistry·1996
Same author

Derivation of a 3D pharmacophore model for the angiotensin-II site one receptor.

Journal of computer-aided molecular design·1994

This study introduces Sample-based Partial Least Squares (SAMPLS), a novel method for analyzing large datasets in molecular modeling. SAMPLS efficiently handles numerous structural properties, improving the prediction of biological activity in drug discovery.

Area of Science:

  • Computational chemistry
  • Cheminformatics
  • Statistical modeling

Background:

  • Three-dimensional molecular modeling generates extensive structural data.
  • Partial Least Squares (PLS) is suitable for large datasets but traditionally property-based.
  • Existing PLS methods are suboptimal for high-dimensional data like Comparative Molecular Field Analysis (CoMFA).

Purpose of the Study:

  • To introduce a sample-based formulation of PLS (SAMPLS) for efficient analysis of large molecular datasets.
  • To enable accurate fitting and prediction of biological responses from complex structural data.
  • To facilitate model validation using modern resampling techniques.

Main Methods:

  • Developed a sample-based PLS formulation (SAMPLS) reducing data to pairwise molecular distances.

Related Experiment Videos

  • Utilized an n-by-n covariance matrix (C) for fitting PLS components.
  • Implemented SAMPLS in Fortran, enabling rapid cross-validation.

    • Main Results:

    • SAMPLS efficiently processes thousands of field values from CoMFA.
    • Full cross-validation of CoMFA data takes only 0.2 CPU seconds.
    • Demonstrated SAMPLS's suitability for structure-activity relationship (SAR) analysis.

    Conclusions:

    • SAMPLS offers a computationally efficient and robust method for PLS analysis in cheminformatics.
    • The sample-distance formulation simplifies information in CoMFA fields and relates PLS to other multivariate methods.
    • SAMPLS is ideal for structure-activity analysis using CoMFA or bonded topology data.