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

Feature selection for DNA methylation based cancer classification.

F Model1, P Adorján, A Olek

  • 1Epigenomics AG, Kastanienallee 24, D-10435 Berlin, Germany. Fabian.Model@epigenomics.com

Bioinformatics (Oxford, England)
|July 27, 2001
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

Epigenomics: genome-wide study of methylation phenomena.

Current issues in molecular biology·2002
Same author

Axonal topography of cortical basket cells in relation to orientation, direction, and ocular dominance maps.

The Journal of comparative neurology·2001
Same author

Active demethylation of the paternal genome in the mouse zygote.

Current biology : CB·2000
Same author

The effect of intracortical competition on the formation of topographic maps in models of Hebbian learning.

Biological cybernetics·2000
Same author

From genomics to epigenomics: a loftier view of life.

Nature biotechnology·1999
Same author

Contrast adaptation and infomax in visual cortical neurons.

Reviews in the neurosciences·1999

Predicting phenotypic classes from high-dimensional molecular data requires careful dimensionality reduction. Combining feature selection and discriminant analysis is crucial for accurate classification, as demonstrated in leukemia subtype discrimination.

Area of Science:

  • Genomics and Bioinformatics
  • Cancer Molecular Profiling
  • Computational Biology

Background:

  • Molecular patterns (mRNA expression, DNA methylation) correlate with phenotypic parameters.
  • Genomic-scale data presents high dimensionality challenges for class prediction.
  • Under-determined problems arise from high data dimensions versus limited samples.

Purpose of the Study:

  • To demonstrate phenotypic class prediction using combined feature selection and discriminant analysis.
  • To evaluate the impact of different dimension reduction strategies on classification performance.
  • To apply these techniques for discriminating between acute lymphoblastic leukemia and acute myeloid leukemia using methylation patterns.

Main Methods:

  • Utilized feature selection techniques to reduce data dimensionality.

Related Experiment Videos

  • Employed discriminant analysis for class prediction.
  • Compared various feature selection methods to identify optimal strategies.

    • Main Results:

    • The choice of dimension reduction strategy significantly impacts classification performance.
    • Combined feature selection and discriminant analysis effectively predict phenotypic classes.
    • Successful discrimination between acute lymphoblastic leukemia and acute myeloid leukemia was achieved using DNA methylation patterns.

    Conclusions:

    • Effective dimensionality reduction is essential for accurate class prediction from high-dimensional molecular data.
    • The proposed method combining feature selection and discriminant analysis offers a robust approach for molecular pattern-based classification.
    • This approach holds promise for improving the diagnosis and understanding of complex diseases like leukemia.