engleski

Prediction of secondary protein structure by means of the electron-ion interaction pseudopotential model

The information necessary for the protein biosynthesis is contained in its messenger RNA nucleotide sequence. The aim of the paper is to provide a model of secondary protein folding based on the nucleotide string analysis, i.e. purine and pyrimidine base recombinations. It is shown that ionisation potential of purine and pyrimidine bases, with respect to their electron donor-acceptor relationships enables the extraction of fast computational algorithms that define secondary protein structure with > 90% precision. The method has been verified by means of the artificial neural network and machine learning procedures, on a test set of 160 alpha- and beta-protein folds from standard protein databases. The algorithms presented in this study enable simple, fast and accurate prediction of the secondary protein folding types on a personal computer.

ionisation potential; machine learning; mRNA; nucleotide; neural network; protein folding

nije evidentirano