The specificity of GalNAc-transferase is consistent with the existence of an extended site composed of nine subsites, denoted by R4, R3, R2, R1, R0, R1', R2', R3', and R4', where the acceptor at R0 is either Ser or Thr to which the reducing monosaccharide is anchored. To predict whether a peptide will react with the enzyme to form a Ser- or Thr-conjugated glycopeptide, a neural network method--Kohonen's self-organization model is proposed in this paper. Three hundred five oligopeptides are chosen for the training site, with another 30 oligopeptides for the test set. Because of its high correct prediction rate (26/30 = 86.7%) and stronger fault-tolerant ability, it is expected that the neural network method can be used as a technique for predicting O-glycosylation and designing effective inhibitors of GalNAc-transferase. It might also be useful for targeting drugs to specific sites in the body and for enzyme replacement therapy for the treatment of genetic disorders.

Our efforts to classify the functional units of many proteins, the modules, are reviewed. The data from the sequencing projects for various model organisms are extremely helpful in deducing the evolution of proteins and modules. For example, a dramatic increase of modular proteins can be observed from yeast to C. elegans in accordance with new protein functions that had to be introduced in multicellular organisms. Our sequence characterization of modules relies on sensitive similarity search algorithms and the collection of multiple sequence alignments for each module. To trace the evolution of modules and to further automate the classification, we have developed a sequence and a module alerting system that checks newly arriving sequence data for the presence of already classified modules. Using these systems, we were able to identify an unexpected similarity between extracellular C1Q modules with bacterial proteins.