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

Profile-based string kernels for remote homology detection and motif extraction.

Rui Kuang1, Eugene Ie, Ke Wang

  • 1Department of Computer Science, Columbia University, 475 Riverside Dr., Mail Code 7717, New York, NY 10115, USA.

Journal of Bioinformatics and Computational Biology
|August 19, 2005
PubMed
Summary

We developed new profile-based string kernels for protein classification and homology detection. These kernels, used with support vector machines (SVMs), significantly outperform existing methods and improve scalability.

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

Author Correction: Ontogeny and transcriptional regulation of Thetis cells.

Nature·2026
Same author

Postmitotic transcription and 3D regulation show locus-specific and differentiation-specific sensitivity to cohesin depletion.

Nature genetics·2026
Same author

Ontogeny and transcriptional regulation of Thetis cells.

Nature·2026
Same author

Robust CD4<sup>+</sup> CAR T cell Expansion Is Associated with Non-ICANS Neurotoxicities Following Ciltacabtagene Autoleucel.

medRxiv : the preprint server for health sciences·2025
Same author

A wave of Thetis cells imparts tolerance to food antigens early in life.

Science (New York, N.Y.)·2025
Same author

Post-mitotic transcriptional activation and 3D regulatory interactions show locus- and differentiation-specific sensitivity to cohesin depletion.

bioRxiv : the preprint server for biology·2025

Area of Science:

  • Bioinformatics
  • Computational Biology
  • Machine Learning

Background:

  • Support vector machines (SVMs) are powerful tools for biological sequence analysis.
  • Existing methods for protein classification and remote homology detection face challenges in scalability and performance.
  • Probabilistic profiles from tools like PSI-BLAST offer rich information about protein sequences.

Purpose of the Study:

  • To introduce novel profile-based string kernels for SVMs.
  • To enhance protein classification and remote homology detection accuracy.
  • To improve the efficiency and scalability of these analyses.

Main Methods:

  • Developed profile-based string kernels utilizing PSI-BLAST probabilistic profiles.
  • Implemented inexact matching of k-mers using position-dependent mutation neighborhoods.

Related Experiment Videos

  • Incorporated predicted secondary structure information into the profile kernel.
  • Extracted discriminative sequence motifs from learned SVM classifiers.

    • Main Results:

    • Profile-based string kernels with SVMs significantly outperformed existing supervised SVM methods in remote homology detection on the SCOP database.
    • Incorporating predicted secondary structure provided a small but significant performance improvement.
    • The method demonstrated superior performance and scalability compared to cluster kernels.
    • Extracted discriminative motifs corresponded to meaningful structural features.

    Conclusions:

    • Profile-based string kernels offer a powerful and efficient approach for protein classification and remote homology detection.
    • The method leverages semi-supervised learning via PSI-BLAST profiles for improved performance.
    • The approach is highly scalable and outperforms previous kernel methods, offering insights into protein structure-function relationships.