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

Borrowing information from relevant microarray studies for sample classification using weighted partial least

Xiaohong Huang1, Wei Pan, Xinqiang Han

  • 1Division of Biostatistics, School of Public Health, University of Minnesota, A460 Mayo Building (MMC 303), Minneapolis, MN 55455-0378, USA.

Computational Biology and Chemistry
|June 28, 2005
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

Using Intervention Mapping to Codevelop Orchid, a Digital Tool for Reproductive Life Planning: Development and Feasibility Study.

JMIR human factors·2026
Same author

Psychological Health, Polysocial Risk, and Cardiovascular Health Among Women of Reproductive Age.

JACC. Advances·2026
Same author

Developing internationally agreed core indicators for surveillance of preconception health: protocol for a consensus study.

PloS one·2026
Same author

Quality of instruments to assess the process of shared decision-making: a comprehensive systematic review and COSMIN quality appraisal.

BMJ evidence-based medicine·2026
Same author

From tokenism to youth leadership in child and adolescent mental health research.

Public health reviews·2026
Same author

Incidence, risk factors and management of paradoxical reaction in central nervous system tuberculosis in the United Kingdom.

The Journal of infection·2026
Same journal

Role of Artificial Intelligence in bioinformatics: Revolutionizing molecular docking and DNA tokenization.

Computational biology and chemistry·2026
Same journal

An interpretable framework for cancer drug response prediction using integrated drug and multi-omics data with a hybrid Bi-LSTM-GRU network.

Computational biology and chemistry·2026
Same journal

SegMWB: A lightweight deep learning framework for microscopic image classification.

Computational biology and chemistry·2026
Same journal

Protein dynamic simulations: From early inception to clinical translation over half a century.

Computational biology and chemistry·2026
Same journal

Integrated omics and virtual screening predict Tabularin as a dual inhibitor of the prognostic microRNAs mir-19a and mir-32 in colorectal cancer.

Computational biology and chemistry·2026
Same journal

In silico characterization of acetyl-CoA carboxylase from Staphylococcus aureus and Escherichia coli: A comparative analysis.

Computational biology and chemistry·2026
See all related articles

This study introduces weighted partial least squares (WPLS) and weighted penalized partial least squares (WPPLS) for improved sample classification using multiple microarray datasets. These methods enhance analysis by effectively borrowing information across studies.

Area of Science:

  • Bioinformatics
  • Computational Biology
  • Statistical Genetics

Background:

  • Publicly available microarray datasets offer opportunities for enhanced data analysis.
  • Traditional meta-analysis assumes data comes from the same population, which is often not the case.
  • Integrating information from multiple, potentially heterogeneous, datasets is crucial for robust sample classification.

Purpose of the Study:

  • To develop novel methods for sample classification using gene expression profiles from multiple microarray datasets.
  • To address the challenge of integrating information from studies with potentially different parameters.
  • To improve classification performance by leveraging relevant external data.

Main Methods:

  • Proposed two new methods: weighted partial least squares (WPLS) and weighted penalized partial least squares (WPPLS).

Related Experiment Videos

  • These methods build classifiers by combining multiple datasets, weighting each dataset by its relevance to the current study.
  • Compared WPLS/WPPLS against a standard approach of combining individual classifiers via weighted voting.

    • Main Results:

    • WPLS/WPPLS significantly improved classification performance compared to methods using only a single dataset.
    • The proposed WPLS/WPPLS methods outperformed the standard approach of combining multiple classifiers.
    • WPPLS demonstrated further improvement over WPLS, analogous to penalized partial least squares (PPLS) over partial least squares (PLS) for single datasets.

    Conclusions:

    • WPLS and WPPLS are effective methods for enhancing sample classification by integrating information from multiple microarray datasets.
    • These methods offer a more powerful alternative to single-dataset analysis and standard classifier combination techniques.
    • The proposed weighted approaches provide a flexible and robust framework for multi-dataset analysis in gene expression studies.