ISMB99 - Tutorial 7

 


Getting the most out of hidden Markov models

Kevin Karplus, Melissa Cline, Christian Barrett

This advanced tutorial will focus on practical issues of building a successful protein structure prediction method from hidden Markov models. The focus of the tutorial is lessons learned in developing SAM-T98, the highest ranked *sequence-based* method in the CASP3 contest. Topics to be covered include tricks for speeding up library searches, construction of a model library, use of sequence weighting and regularizers to improve model generality, null model selection, and significance levels of null model scores. Participants are expected to have a basic knowledge of profile HMMs; background material will be provided in the handouts.

-> Tutorial Handouts
-> Tutorial Program
-> ISMB 99