Related Experiment Videos
Approximate p-values for local sequence alignments: numerical studies.
1Department of Statistics, Stanford University, Stanford, CA 94305, USA.
Summary
This study simplifies calculating approximate p-values for sequence alignment. The new method reduces complex parameters to two computable values, improving accuracy for general scoring matrices and affine gap penalties.
Area of Science:
- Bioinformatics
- Computational Biology
- Statistical Genetics
Background:
- Optimal sequence alignment is crucial for understanding biological relationships.
- Existing p-value approximations involve complex, computationally intensive parameters.
- General scoring matrices and affine gap penalties are widely used in sequence analysis.
Purpose of the Study:
- To simplify and improve the accuracy of approximate p-value calculations in sequence alignment.
- To reduce the computational complexity associated with existing approximation methods.
- To provide a more accessible method for evaluating alignment significance.
Main Methods:
- Numerical studies were conducted to analyze the parameters of an existing approximation.
- The infinite sequence of parameters was shown to reduce to two key values.
- These parameters were computed using one-dimensional numerical integrals.
Main Results:
- The complex approximation parameters were found to simplify to essentially two distinct numerical values.
- A modified approximation was developed using these two parameters.
- This modified approximation is easily evaluated for arbitrary scoring matrices and affine gap penalties.
Conclusions:
- The simplified approximation offers a computationally efficient and accurate method for estimating p-values in sequence alignment.
- The findings facilitate more accessible statistical significance testing in bioinformatics.
- The approach is robust for various scoring systems and gap penalty models.