Header logo is

A simple iterative approach to parameter optimization

2000

Conference Paper

ei


Various bioinformatics problems require optimizing several different properties simultaneously. For example, in the protein threading problem, a linear scoring function combines the values for different properties of possible sequence-to-structure alignments into a single score to allow for unambigous optimization. In this context, an essential question is how each property should be weighted. As the native structures are known for some sequences, the implied partial ordering on optimal alignments may be used to adjust the weights. To resolve the arising interdependence of weights and computed solutions, we propose a novel approach: iterating the computation of solutions (here: threading alignments) given the weights and the estimation of optimal weights of the scoring function given these solutions via a systematic calibration method. We show that this procedure converges to structurally meaningful weights, that also lead to significantly improved performance on comprehensive test data sets as measured in different ways. The latter indicates that the performance of threading can be improved in general.

Author(s): Zien, A. and Zimmer, R. and Lengauer, T.
Book Title: RECOMB2000
Journal: Proceedings of the Forth Annual Conference on Research in Computational Molecular Biology (RECOMB2000)
Pages: 318-327
Year: 2000
Month: April
Day: 0
Publisher: ACM Press

Department(s): Empirical Inference
Bibtex Type: Conference Paper (inproceedings)

DOI: 10.1145/332306.332570
Event Name: Forth Annual Conference on Research in Computational Molecular Biology
Event Place: Tokyo, Japan

Address: New York, NY, USA
Digital: 0
Institution: Fraunhofer Institute SCAI
Language: en
Organization: Max-Planck-Gesellschaft
School: Biologische Kybernetik

Links: Web

BibTex

@inproceedings{2139,
  title = {A simple iterative approach to parameter optimization},
  author = {Zien, A. and Zimmer, R. and Lengauer, T.},
  journal = {Proceedings of the Forth Annual Conference on Research in Computational Molecular Biology (RECOMB2000)},
  booktitle = {RECOMB2000},
  pages = {318-327},
  publisher = {ACM Press},
  organization = {Max-Planck-Gesellschaft},
  institution = {Fraunhofer Institute SCAI},
  school = {Biologische Kybernetik},
  address = {New York, NY, USA},
  month = apr,
  year = {2000},
  month_numeric = {4}
}