Improving Sequence Alignment Statistics with Position-Dependent Scoring
Author Information
Author(s): Stefan Wolfsheimer, Inke Herms, Sven Rahmann, Alexander K. Hartmann
Primary Institution: Laboratoire MAP5 (UMR CNRS 8145), Université Paris Descartes, France
Hypothesis
Can we develop a method to accurately compute score distributions for sequence alignments that accounts for position-dependent scoring?
Conclusion
The study presents a new method that significantly improves the accuracy of score distributions for sequence alignments, particularly for transmembrane proteins.
Supporting Evidence
- The method allows access to the low-probability region of the score distribution.
- It demonstrates improved sensitivity and specificity for detecting homologous transmembrane proteins.
- The results indicate that position-dependent scoring significantly affects alignment statistics.
Takeaway
This study helps scientists better understand how similar DNA or protein sequences are by using a new method that looks at how the position of each part of the sequence matters.
Methodology
The study uses Markov chain Monte Carlo simulations combined with importance sampling to estimate score distributions for various sequence comparison models.
Limitations
The method currently requires a high number of samples, making it impractical for real-time database searches.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website