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

Classification of a large microarray data set: algorithm comparison and analysis of drug signatures.

Georges Natsoulis1, Laurent El Ghaoui, Gert R G Lanckriet

  • 1Iconix Pharmaceuticals, Mountain View, CA 94043, USA. gnatsoulis@iconixpharm.com

Genome Research
|May 4, 2005
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

The lysine-specific demethylase 1 (LSD1) inhibitor bomedemstat in myelofibrosis: results from a phase 1/2 study.

Blood advances·2026
Same author

Phase 2 study of the lysine-specific demethylase 1 inhibitor bomedemstat for essential thrombocythemia.

Blood advances·2026
Same author

A new Jun amino-terminal kinase inhibitor, KRev-202, inhibits rat ischemic acute injury and the progression to renal fibrosis.

Frontiers in pharmacology·2026
Same author

The Influence of Radiographic Equipment, Setup, and Operator Experience on Radiation Exposure in Chronic Total Occlusion Interventions.

Catheterization and cardiovascular interventions : official journal of the Society for Cardiac Angiography & Interventions·2025
Same author

Long-term predictors of target vessel failure after intracoronary lithotripsy: 12-month results from the France LILI registry.

EuroIntervention : journal of EuroPCR in collaboration with the Working Group on Interventional Cardiology of the European Society of Cardiology·2025
Same author

A humanised ACE2, TMPRSS2, and FCGRT mouse model reveals the protective efficacy of anti-receptor binding domain antibodies elicited by SARS-CoV-2 hybrid immunity.

EBioMedicine·2025

Researchers developed interpretable drug signatures using machine learning on gene expression data. Support Vector Machines and Logistic Regression identified key genes, aiding biomarker discovery.

Area of Science:

  • Toxicogenomics
  • Bioinformatics
  • Computational Biology

Background:

  • A comprehensive gene expression database details drug and toxicant effects in rats.
  • Extracting biological insights from large-scale toxicogenomic data presents a significant challenge.

Purpose of the Study:

  • To evaluate supervised classification algorithms for deriving interpretable drug signatures from a large gene expression database.
  • To identify robust methods for analyzing drug-induced gene expression patterns and developing diagnostic biomarkers.

Main Methods:

  • Utilized a 597-microarray subset of the gene expression database.
  • Compared various supervised classification algorithms, focusing on Support Vector Machines (SVMs) and Logistic Regression.
  • Employed feature selection techniques to refine drug signatures into short, weighted gene lists.

Related Experiment Videos

Main Results:

  • Linear classifiers, particularly SVMs and Logistic Regression, achieved high classification performance.
  • Identified interpretable drug signatures composed of 'reward' and 'penalty' genes, enhancing classification accuracy and reducing false positives.
  • Feature selection successfully reduced signature length, facilitating biomarker development.

Conclusions:

  • Supervised classification algorithms, especially SVMs and Logistic Regression, are effective for generating interpretable drug signatures from gene expression data.
  • The developed signatures, combining reward and penalty genes, offer a promising approach for diagnostic biomarker and low-cost assay development.
  • Comparing multiple signatures reveals characteristic biological processes associated with specific drug classes.