A comparative study of S/MAR prediction tools
2007

Evaluating S/MAR Prediction Tools

Sample size: 165 publication Evidence: low

Author Information

Author(s): Evans Kenneth, Ott Sascha, Hansen Annika, Koentges Georgy, Wernisch Lorenz

Primary Institution: Birkbeck College, University College London

Hypothesis

How accurately can existing S/MAR prediction methods identify scaffold/matrix attachment regions in DNA?

Conclusion

Existing S/MAR prediction methods have little predictive power, and a simple rule based on AT-percentage performs comparably.

Supporting Evidence

  • All existing S/MAR prediction methods have little predictive power.
  • A simple rule based on AT-percentage is competitive with other methods.
  • Different methods identify different sub-sequences as S/MARs.
  • Further research on the H-Rule is suggested for better predictions.

Takeaway

Scientists tried different computer methods to find special DNA regions called S/MARs, but none worked very well, and a simple rule based on the amount of AT in the DNA was just as good.

Methodology

The study evaluated various S/MAR prediction methods using a positive test set of known S/MARs and control datasets to measure their predictive power.

Potential Biases

Previous analyses may have been biased by focusing on a few best cases and not including adequate control comparisons.

Limitations

The methods analyzed could not serve as practical prediction tools, and the dataset size may limit the generalizability of the findings.

Digital Object Identifier (DOI)

10.1186/1471-2105-8-71

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication