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

Global discriminative learning for higher-accuracy computational gene prediction.

Axel Bernal1, Koby Crammer, Artemis Hatzigeorgiou

  • 1Department of Computer and Information Science, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America. abernal@seas.upenn.edu

Plos Computational Biology
|March 21, 2007
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

Stability in the face of global decline: a 20-year study of arthropods in an oceanic archipelago.

Scientific reports·2026
Same author

Use of Phosphorus-32 Microparticles for Downstaging Locally Advanced Pancreatic Cancer to Resection: A Case Report.

Cureus·2026
Same author

Liver disease with extrahepatic peritoneal disease.

Cirugia espanola·2026
Same author

Integrating plot-based methods for monitoring biodiversity in island habitats under the scope of BIODIVERSA+ project BioMonI: Beetle monitoring in Pico and Terceira Islands.

Biodiversity data journal·2026
Same author

New spatial records of vascular plants in the Azores Archipelago: the PRIBES project and the Azorean Biodiversity Portal (ABP) initiatives - I. São Jorge Island (Azores).

Biodiversity data journal·2026
Same author

Comparative clinical trial of open versus closed technique for intraoperative hyperthermic intraperitoneal chemotherapy in peritoneal metastases from colorectal cancer.

Current problems in surgery·2026

This study introduces CRAIG, a novel ab initio gene prediction tool. CRAIG utilizes a discriminative learning approach for improved accuracy in gene annotation, especially for complex genomic regions.

Area of Science:

  • Computational Biology
  • Bioinformatics
  • Genomics

Background:

  • Most ab initio gene predictors rely on hidden Markov models, which have limitations in optimizing prediction accuracy and handling statistical dependencies.
  • Piecewise training of genomic signals and content separately hinders overall gene model accuracy.

Purpose of the Study:

  • To investigate alternative approaches for ab initio gene prediction by integrating diverse genomic evidence.
  • To develop a novel gene predictor that maximizes annotation accuracy using discriminative learning.

Main Methods:

  • Developed CRAIG, a gene prediction program employing a conditional random field model with a semi-Markov structure.
  • Trained the model using an online large-margin algorithm, related to multiclass support vector machines (SVMs).

Related Experiment Videos

  • Integrated multiple types of genomic evidence with complex statistical dependencies.

    • Main Results:

    • CRAIG demonstrated significant improvements in prediction accuracy compared to existing intrinsic-feature-only gene predictors.
    • Accuracy gains were particularly notable at the gene level and for genes with long introns.
    • Experiments were conducted on benchmark vertebrate datasets and ENCODE project regions.

    Conclusions:

    • Discriminative learning offers a powerful approach to enhance ab initio gene prediction accuracy.
    • CRAIG provides a more accurate method for gene annotation, especially in complex genomic contexts.
    • The findings highlight the potential of integrating diverse genomic data for improved gene prediction.