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

Recursive SVM feature selection and sample classification for mass-spectrometry and microarray data.

Xuegong Zhang1, Xin Lu, Qian Shi

  • 1Bioinformatics Div, TNLIST, Dept of Automation, Tsinghua University, Beijing, 100084, China. xgzhang@tsinghua.edu.cn

BMC Bioinformatics
|April 12, 2006
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

Perspectives From an Expert-Guided Discussion on Maximizing the Research Potential of Small Biopsy Tissue.

JCO precision oncology·2026
Same author

JADE: Joint Alignment and Deep Embedding for Multi-Slice Spatial Transcriptomics.

Advances in neural information processing systems·2026
Same author

Phase II Study of Adavosertib in Patients With Tumors Containing BRCA1 and BRCA2 Mutations: Results From the NCI-MATCH ECOG-ACRIN Cancer Research Group (EAY131) Subprotocol Z1I.

JCO precision oncology·2026
Same author

Transformative microsystems for Cancer Diagnosis, treatment monitoring, and clinical integration: Outcomes from an NCI Workshop.

Journal of the National Cancer Institute·2026
Same author

Artificial intelligence in cancer diagnostics: a primer for clinicians and scientists.

Journal of the National Cancer Institute·2026
Same author

CD133-Guided RNA Nanoparticle Delivery of FTO siRNA Impairs Leukemia Resistance to Tyrosine Kinase Inhibitor Therapy.

RNA nanomed·2026

A new recursive support vector machine (R-SVM) algorithm improves biomarker discovery in noisy high-throughput proteomics and microarray data. R-SVM demonstrates superior robustness and feature recovery compared to existing methods like SVM recursive feature elimination (SVM-RFE).

Area of Science:

  • Bioinformatics
  • Computational Biology
  • Proteomics

Background:

  • High-throughput omics data, including proteomics and microarrays, necessitate robust machine learning for biomarker identification.
  • Effective feature selection and classification algorithms must handle noise and outliers inherent in biological datasets.

Purpose of the Study:

  • To develop and evaluate a novel recursive support vector machine (R-SVM) algorithm for enhanced biomarker discovery.
  • To assess the performance of R-SVM against state-of-the-art methods, specifically SVM recursive feature elimination (SVM-RFE), in terms of feature recovery and robustness.

Main Methods:

  • Development of a recursive support vector machine (R-SVM) algorithm for feature selection in noisy biological data.
  • Comparative analysis of R-SVM and SVM-RFE using simulation experiments and real-world proteomics datasets (human breast cancer, rat liver cirrhosis).

Related Experiment Videos

  • Validation of identified biomarkers through subsequent biological experiments.

    • Main Results:

    • R-SVM achieved a 5%-20% improvement over SVM-RFE in recovering informative biomarkers and robustness to data outliers.
    • Support Vector Machine (SVM)-based methods outperformed conventional univariate methods in classification accuracy.
    • Univariate methods showed strengths in identifying differentially expressed features, particularly with correlated data.

    Conclusions:

    • The R-SVM method is highly effective for analyzing noisy high-throughput omics data, outperforming SVM-RFE.
    • Multivariate SVM approaches offer superior classification performance compared to univariate methods.
    • Understanding the trade-offs between multivariate and univariate methods is crucial for comprehensive feature analysis in omics studies.