engleski

Prediction of protein folding types by means of Symbolic Cantor Algorithm.

We present new algorithm that enables the evaluation of protein folding types from the nucleotide and amino acid sequences. The method has been denoted Symbolic Cantor Algorithm, i.e. SCA (Štambuk N: Croat Chem Acta 71, 573, 1998; Period Biol, 101, 355, 1999). The algorithm was compared to other methods of protein prediction based on hydropathic indexes, amino acid cloister energies, molecular volumes, polarities and molecular resonant recognition method. Our results indicate that SCA shows considerable improvement in the accuracy and specificity of the secondary structure classification. Statistical and neural network testing of the model validity was performed on different proteins from the relevant databases. SCA procedure of computational analysis does not require the use of computers with large processing capacity and it performs almost immediate neural netwotk based pattern recognition of the secondary protein structure.

protein folding; computation; algorithm; prediction; neural network

nije evidentirano