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

Neural network input representations that produce accurate consensus sequences from DNA fragment assemblies.

C F Allex1, J W Shavlik, F R Blattner

  • 1Computer Sciences Department, University of Wisconsin - Madison, 1210 West Dayton Street, Madison, WI 53706, DNASTAR Inc., 1228 South Park Street, Madison, WI 53715, USA.

Bioinformatics (Oxford, England)
|September 28, 1999
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

Biomedical informatics training at the University of Wisconsin-Madison.

Yearbook of medical informatics·2007
Same author

Sex-specific SCAR markers in the dioecious plant Rumex nivalis (Polygonaceae) and implications for the evolution of sex chromosomes.

TAG. Theoretical and applied genetics. Theoretische und angewandte Genetik·2003
Same author

Storage and mobilization as antagonistic functional constraints on seed storage globulin evolution.

Journal of experimental botany·2003
Same author

Complete genome sequence and comparative genomics of Shigella flexneri serotype 2a strain 2457T.

Infection and immunity·2003
Same author

Extensive mosaic structure revealed by the complete genome sequence of uropathogenic Escherichia coli.

Proceedings of the National Academy of Sciences of the United States of America·2002
Same author

Nunatak survival of the high Alpine plant Eritrichium nanum (L.) Gaudin in the central Alps during the ice ages.

Molecular ecology·2002

Neural networks trained with DNA sequencing trace data achieve high accuracy in determining consensus bases. Incorporating trace information significantly improves base call accuracy compared to using base calls alone.

Area of Science:

  • Genomics
  • Bioinformatics
  • Computational Biology

Background:

  • DNA sequencing generates fluorescent traces and base calls for aligned DNA sequences.
  • Accurate consensus base determination is crucial for reliable genomic data analysis.
  • Developing robust computational methods for DNA base calling is an ongoing challenge.

Purpose of the Study:

  • To compare different input representations for neural networks in DNA base calling.
  • To evaluate the impact of trace information versus base calls alone on prediction accuracy.
  • To optimize neural network performance for accurate consensus base determination.

Main Methods:

  • Empirical comparison of five distinct input representations for neural networks.
  • Training neural networks using aligned DNA base calls and Perkin Elmer Applied Biosystems (ABI) fluorescent traces.

Related Experiment Videos

  • Utilizing 10-fold cross-validation to estimate network performance and accuracy.

    • Main Results:

    • Neural networks incorporating trace information achieved superior accuracy in consensus base calling.
    • The best network topology yielded consensus accuracies between 99.26% and >99.98% for varying sequence coverages.
    • Networks using only base calls exhibited a higher error rate (8 in 20,000) compared to those using trace data (3 in 20,000).

    Conclusions:

    • Trace information is a critical feature for improving neural network-based DNA base calling.
    • The developed neural network models offer a significant advancement in the accuracy of consensus base determination.
    • These findings have implications for enhancing the reliability of high-throughput sequencing data analysis.