Analyzing Protein Data with the Generative Topographic Mapping Approach
The Generative Topographic Mapping (GTM) approach by Bishop et al. (1998) is used for the classification of sequences from a protein family and the graphical display of the group relationships on a two-dimensional map. The results are compared with the closely related Self-Organizing Map (SOM) approach of Kohonen (1982). A modification of the classical GTM approach is presented, better suited for the analysis of sequence data.
KeywordsLatent Space Data Space Parameter Matrix Alignment Position Related Amino Acid
Unable to display preview. Download preview PDF.
- GRIMMENSTEIN, I.M., ANDRADE, M.A., and URFER, W. (2002): Identification of Conserved Regions in Protein Families by Self-Organizing Maps. Technical Report 36/2002, SFB 475, Department of Statistics, University of Dortmund.Google Scholar
- SVENSÉN, M. (1998): GTM: The Generative Topographic Mapping. PhD Thesis, Aston University.Google Scholar