PROTEIN SECONDARY STRUCTURE PREDICTION USING LOCAL ALIGNMENTS

The accuracy of secondary structure prediction methods has been improv ed significantly by the use of aligned protein sequences. The PHD meth od and the NNSSP method reach 71 to 72% of sustained overall three-sta te accuracy when multiple sequence alignments are with neural networks and nearest-neighbor algorithms, respectively. We introduce a variant of the nearest-neighbor approach that can achieve similar accuracy us ing a single sequence as the query input. We compute the 50 best non-i ntersecting local alignments of the query sequence with each sequence from a set of proteins with known 3D structures. Each position of the query sequence is aligned with the database amino acids in alpha-helic al, beta-strand or coil states. The prediction type of secondary struc ture is selected as the type of aligned position with the maximal tota l score. On the dataset of 124 non-membrane non-homologous proteins, u sed earlier as a benchmark for secondary structure predictions, our me thod reaches an overall three-state accuracy of 71.2%. The performance accuracy is verified by an additional test on 461 non-homologous prot eins giving an accuracy of 71.0%. The main strength of the method is t he high level of prediction accuracy for proteins without any known ho molog. Using multiple sequence alignments as input the method has a pr ediction accuracy of 73.5%. Prediction of secondary structure by the S SPAL method is available via Baylor College of Medicine World Wide Web server. (C) 1997 Academic Press Limited.