PREDICTION OF THE SECONDARY STRUCTURE OF GLOBULAR-PROTEINS BY THEIR AMINO-ACID-SEQUENCE

Prediction algorithms for protein secondary structures by amino acid sequence based on a statistical analysis of 46 proteins were devised. The ith residual chain is assumed to be in 1 of 4 states: .alpha.-helix, .beta.-sheet, .beta.-turn or random coil conformation. The algorithm enables proper prediction of 85.9% of the residual states in .alpha.-helical conformation, 72.4% in .beta.-sheet conformation and 70.7% in .beta.-turn conformation. A comparison was made with other prediction methods of the protein secondary structure, which showed that the algorithms proposed are the most effective ones.