Improved benchmarks for computational motif discovery
2007

Improved benchmarks for computational motif discovery

publication Evidence: moderate

Author Information

Author(s): Sandve Geir Kjetil, Abul Osman, Walseng Vegard, Drabløs Finn

Primary Institution: Norwegian University of Science and Technology (NTNU)

Hypothesis

How can we improve the assessment of motif discovery methods for better evaluation of their performance?

Conclusion

The new benchmark suites allow for better differentiation between the performance of motif discovery algorithms and the power of motif models.

Supporting Evidence

  • The study highlights the difficulty of accurately assessing motif discovery methods due to the complexity of biological data.
  • New benchmark suites were created to provide clearer evaluations of motif discovery algorithms.
  • The research indicates that common motif models may not always effectively discriminate between positive and negative instances.

Takeaway

This study created better ways to test how well tools can find important patterns in DNA, helping scientists choose the best tools for their research.

Methodology

The study used machine learning algorithms to analyze transcription factor binding sites and developed new benchmark suites for evaluating motif discovery methods.

Potential Biases

Synthetic data sets may introduce bias towards specific classes of tools, affecting the generalizability of the results.

Limitations

The initial alignment of binding sites may be sub-optimal, and the performance of motif discovery methods can be influenced by the complexity of the data sets.

Statistical Information

P-Value

p<0.05

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1186/1471-2105-8-193

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication