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

Unsupervised classification of noisy chromosomes.

T Y Chan1

  • 1The University of Aizu, Aizu-Wakamatsu Shi, Fukushima Ken, 965-80 Japan. t-chan@u-aizu.ac.jp

Bioinformatics (Oxford, England)
|May 2, 2001
PubMed
Summary

This study introduces an unsupervised method for chromosome recognition, enabling systems to distinguish between chromosome types without prior labels. The approach dynamically learns classification weights, effectively handling noisy 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

A Material Approach to Endangered Species Conservation: Characterization and 3D Imaging of Ballistic Damage in the Casques of Helmeted Hornbill (<i>Rhinoplax vigil</i>).

Integrative organismal biology (Oxford, England)·2026
Same author

Implications for patient follow-up of changes in inferior vena cava filter removal plan practices.

Clinical radiology·2022
Same author

Variability in the perception and application of force used in IVC filter retrievals among interventional radiologists.

Radiography (London, England : 1995)·2021
Same author

Porcine dermal collagen mesh (Permacol™) as a bioprosthesis in the ligation of intersphincteric tract (BioLIFT) procedure.

Techniques in coloproctology·2020
Same author

The dynamic and stress-adaptive signaling hub of 14-3-3: emerging mechanisms of regulation and context-dependent protein-protein interactions.

Oncogene·2018
Same author

Evidence of malignant hyperthermia in patients administered triggering agents before malignant hyperthermia susceptibility identified: missed opportunities prior to diagnosis.

Anaesthesia and intensive care·2017

Area of Science:

  • Computational Biology
  • Machine Learning
  • Genetics

Background:

  • Traditional chromosome recognition relies on supervised learning with pre-classified data.
  • Existing methods are sensitive to noise in chromosome representation but assume correct initial classification.
  • Recognition typically involves calculating string edit distance to known chromosome representatives.

Purpose of the Study:

  • To develop a general method for unsupervised chromosome induction and recognition.
  • To address the challenge of classifying chromosomes without prior knowledge of their labels.
  • To enable dynamic learning of classification criteria in the presence of noisy data.

Main Methods:

  • Introduced a novel unsupervised learning framework for chromosome recognition.
  • Developed an inductive agent capable of learning classification weights dynamically.
  • Characterized the learning process as finding an optimal distance function for class separation.

Main Results:

  • Successfully demonstrated unsupervised distinction between noisy median and telocentric chromosomes.
  • The inductive agent learned to differentiate chromosome types without explicit labels.
  • The method effectively identifies the appropriate distance function for accurate classification.

Conclusions:

  • The proposed unsupervised method offers a robust solution for chromosome recognition, particularly with noisy or unlabeled data.
  • Dynamic weight learning allows for adaptive classification without predefined labels.
  • This approach advances machine learning applications in cytogenetics and bioinformatics.

Related Experiment Videos