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

Recovering genetic regulatory networks from micro-array data and location analysis data.

Fan Li1, Yiming Yang

  • 1LTI, School of Computer Science, Carnegie Mellon Univ., 4502 Newell Simon Hall, 5000 Forbes Ave, Pittsburgh, PA 15213, USA. hustlf@cs.cmu.edu

Genome Informatics. International Conference on Genome Informatics
|February 12, 2005
PubMed
Summary

This study addresses challenges in learning large biological networks from high-throughput data. A new Lasso regression-based algorithm significantly outperforms existing methods, especially with unbalanced variable-to-instance ratios common in microarray data.

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

Dynamics of male meiotic recombination frequency during plant development using Fluorescent Tagged Lines in Arabidopsis thaliana.

Scientific reports·2017
Same author

Regenerative Polysulfide-Scavenging Layers Enabling Lithium-Sulfur Batteries with High Energy Density and Prolonged Cycling Life.

ACS nano·2017
Same author

PdAuCu Nanobranch as Self-Repairing Electrocatalyst for Oxygen Reduction Reaction.

ChemSusChem·2017
Same author

Trapdoor spiders of the genus <i>Cyclocosmia</i> Ausserer, 1871 from China and Vietnam (Araneae, Ctenizidae).

ZooKeys·2017
Same author

The complete genome sequence, occurrence and host range of Tomato mottle mosaic virus Chinese isolate.

Virology journal·2017
Same author

Tunneling nanotubes promote intercellular mitochondria transfer followed by increased invasiveness in bladder cancer cells.

Oncotarget·2017

Area of Science:

  • Computational Biology
  • Bioinformatics
  • Systems Biology

Background:

  • High-throughput biological data, like micro-array data, necessitates learning large-scale networks.
  • Existing algorithms like Sparse Candidate Hill Climbing (SCHC) and Grow-Shrinkage (GS) face effectiveness issues with large numbers of variables relative to instances.

Purpose of the Study:

  • To investigate the performance limitations of SCHC and GS algorithms for large-scale network learning.
  • To propose and evaluate a novel network structure learning algorithm utilizing Lasso regression for improved performance, particularly in unbalanced data scenarios.

Main Methods:

  • Comparative analysis of SCHC and GS algorithms on synthetic and real biological data.
  • Development of a new large-scale network structure learning algorithm based on Lasso regression.

Related Experiment Videos

  • Validation using synthetic datasets and a real micro-array dataset (over 6000 genes) combined with genome-wide location analysis data.

    • Main Results:

    • SCHC and GS algorithms exhibit significant effectiveness problems when the number of variables greatly exceeds the number of instances.
    • The proposed Lasso regression-based algorithm demonstrates superior performance compared to SCHC and GS on synthetic data.
    • The new algorithm successfully identified biologically relevant genetic regulatory network modules from real micro-array data.

    Conclusions:

    • The proposed Lasso regression-based algorithm is highly effective for large-scale network structure learning, especially with unbalanced variable-to-instance ratios.
    • The algorithm's ability to learn meaningful biological networks from complex datasets is validated.
    • This approach offers a promising solution for analyzing high-throughput biological data to uncover gene regulatory mechanisms.