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

Hidden Markov models that use predicted secondary structures for fold recognition.

J Hargbo1, A Elofsson

  • 1Department of Biochemistry, Stockholm University, Sweden.

Proteins
|June 18, 1999
PubMed
Summary

Protein fold recognition methods improve with predicted secondary structures and multiple sequence alignments. Combining these approaches correctly identifies protein folds in 20% of cases, outperforming standard sequence search methods.

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

pyconsFold: a fast and easy tool for modeling and docking using distance predictions.

Bioinformatics (Oxford, England)·2021
Same author

The SubCons webserver: A user friendly web interface for state-of-the-art subcellular localization prediction.

Protein science : a publication of the Protein Society·2017
Same author

SubCons: a new ensemble method for improved human subcellular localization predictions.

Bioinformatics (Oxford, England)·2017
Same author

CAFASP2: the second critical assessment of fully automated structure prediction methods.

Proteins·2002
Same author

LiveBench-2: large-scale automated evaluation of protein structure prediction servers.

Proteins·2002
Same author

Pcons: a neural-network-based consensus predictor that improves fold recognition.

Protein science : a publication of the Protein Society·2001

Area of Science:

  • Computational biology
  • Structural bioinformatics
  • Protein structure prediction

Background:

  • Many proteins share similar structures (folds) despite lacking sequence similarity.
  • Protein fold recognition methods aim to predict these structures.
  • Recent advancements utilize predicted secondary structures and hidden Markov models (HMMs).

Purpose of the Study:

  • To rigorously benchmark and compare different protein fold recognition methods.
  • To evaluate the performance of HMMs against standard sequence search methods.
  • To assess the combined efficacy of predicted secondary structures and multiple sequence alignments.

Main Methods:

  • Development of a comprehensive benchmark dataset for protein fold recognition.
  • Comparison of automatically generated HMMs with standard sequence search algorithms (e.g., BLAST, FASTA).

Related Experiment Videos

  • Implementation and evaluation of a combined method integrating predicted secondary structures and multiple sequence alignments.

    • Main Results:

    • The combined method, using predicted secondary structures and multiple sequence alignments, achieved 20% accuracy in fold identification.
    • Using only single sequences yielded 10% accuracy.
    • Including multiple sequence information improved accuracy to 16%.
    • With correct secondary structure information, accuracy reached 27%.
    • Standard methods (blast2, fasta, ssearch) achieved 13-17% accuracy.
    • HMMs performed comparably to standard methods, potentially due to limited sequence diversity in alignments and HMM generation limitations.

    Conclusions:

    • Combining predicted secondary structures and multiple sequence alignments significantly enhances protein fold recognition.
    • Current HMMs may require further refinement for optimal performance, especially with limited sequence data.
    • The developed benchmark provides a valuable resource for evaluating future protein structure prediction tools.