USING NEURAL-NETWORK PREDICTED SECONDARY STRUCTURE INFORMATION IN AUTOMATIC PROTEIN NMR ASSIGNMENT

In CAPRI, an automated NMR assignment software package that was develo ped in our laboratory, both chemical shift values and coupling topolog ies of spin patterns are used ina procedure for amino acids recognitio n. By using a knowledge base of chemical shift distributions of the 20 amino acid types, fuzzy mathematics, and pattern recognition theory, the spin coupling topological graphs are mapped onto specific amino ac id residues. In this work, we investigated the feasibility of using se condary structure: information of proteins as predicted by neural netw orks in the automated NMR assignment, As the H-1 and C-13 chemical shi fts of proteins are known to correlate to their secondary structures, secondary structure information is useful in improving the amino acid recognition, in this study, the secondary structures of proteins predi cted by the PHD protein server and our own trained neural networks are used in the amino acid type recognition, The results show that the pr edicted secondary structure information can help to improve the accura cy of the amino acid recognition.