Issue No. 06 - November/December (2001 vol. 16)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/5254.972074
<p>This article describes an implementation of a set of rules for automatically embedding a minimal functional sequence pattern in a structural hidden Markov model. An HMM is constructed from the atomic coordinates of a protein structure deposited in the PDB (protein databank) by an entirely automated procedure. This procedure generates a set of HMMs that represent distinct three-dimensional folds of protein structures. The next step generates a sequence pattern for the functional family from a set of homologous sequences. The strictly conserved positions from the pattern are selected for embedding in each structural HMM. The final product is a library of wide-ranging fold models with encoded information about the functional families. This library is used to assign the fold models to protein sequences.</p>
J. Bienkowska, T. F. Smith and H. He, "Automatic Pattern Embedding in Protein Structure Models," in IEEE Intelligent Systems, vol. 16, no. , pp. 21-25, 2001.