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

A collaborative filtering-based approach to biomedical knowledge discovery.

Jake Lever1,2, Sitanshu Gakkhar1, Michael Gottlieb1

  • 1Canada's Michael Smith Genome Sciences Centre, Vancouver, BC V5Z 4S6, Canada.

Bioinformatics (Oxford, England)
|October 14, 2017
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

Experimental and computational approaches for deep metabolome annotation with application to the ecotoxicological model organism Daphnia magna.

GigaScience·2026
Same author

Double robustness.

Nature methods·2026
Same author

Estrogen Exposure is Associated With Reduced Otosclerosis Risk in Obesity and Hormone Therapy.

Otology & neurotology open·2026
Same author

The Typability Index: A tool for measuring and controlling for typing difficulty in text stimuli.

Behavior research methods·2026
Same author

CIViC MCP: Integrating Large Language Models with the Clinical Interpretations of Variants in Cancer.

bioRxiv : the preprint server for biology·2025
Same author

CeRTS: certainty retrieval token search in large language model clinical information extraction.

Journal of biomedical informatics·2025

Singular value decomposition (SVD) advances literature-based discovery by using global knowledge graph data. This approach outperforms existing methods in predicting future biomedical associations from published research.

Area of Science:

  • Biomedical Informatics
  • Computational Biology
  • Knowledge Discovery

Background:

  • Increasing publication rates challenge researchers in identifying novel hypotheses.
  • Literature-based discovery methods use knowledge graphs to predict future biomedical associations.
  • Current prediction methods rely on local graph structures, limiting their scope.

Purpose of the Study:

  • To develop a novel method for literature-based discovery using global knowledge graph data.
  • To improve the prediction of future biomedical associations for research hypothesis generation.

Main Methods:

  • Propose a singular value decomposition (SVD) based approach for literature-based discovery.
  • Utilize co-occurrence data from published literature to build and analyze knowledge graphs.

Related Experiment Videos

  • Employ a reduced representation to integrate data across the entire knowledge graph.

    • Main Results:

    • The SVD approach demonstrates superior performance compared to leading methods in scoring discoveries.
    • Predictive power diminishes for associations appearing further in the future.
    • Comparative analysis highlights the strengths and weaknesses of SVD against other systems.

    Conclusions:

    • SVD offers a robust method for literature-based discovery by leveraging global knowledge graph information.
    • This approach enhances the ability of researchers to identify novel research hypotheses.
    • Knowledge discovery tools can become more frequently utilized with advanced prediction methods.