Prediction Of Protein Folding Rates Based On Multi Feature Fusion Parameters

Protein can rely on the interaction assembling their own from in the cellular environment from an unstable state to the stable state. The self assembling process is called protein folding.Only correctly folded proteins into their native structures in particular,can It perform its biological function.Therefore,the correct understanding of the mechanism of protein folding is a core subject of Modern Biophysics,It not only has important scientific significance,but also has great application value in the field of medical and biological engineering.To determine the influence factors of protein folding rate has important value for the deep exploration of the protein folding mechanism and principle.In recent years,many researchers proposed various characteristic parameters of protein folding rate and some protein folding rate prediction method based on these characteristic parameters.This paper puts forward two kinds of protein folding rate prediction method.The main work is as follows:Just using the Pseudo-acid composition which provided by Chou et al,we provide a new pseudo-acid composition from the amino acid composition information and the order of the protein sequence.It is that combining the autocorrelation function with the Nm and Frequency of amino acids to Construct 23-dimensional vector so that a protein sequence can be described by the 23-dimensional vector.And then creating multiple linear regression function to predicte protein folding rate.Ander the jackknife text the correlation coefficient is 0.84.We compare our method with the other two methods to make the conclusion that ours is batter.Combining four protein sequences component information.They are the‘amino acid frequences’ ‘protein sequences of LZ complexity’‘amino acid polarity correlation coefficient’‘amino acid index correlation coefficient’.Then using the four parameters to describe the process of protein folding.Between the feature vectors and the protein folding rates experiment values establish multiple linear regression function.After that we get prediction values.Ander the Jackknife text the correlation coefficient is 0.89.It demonstrate that these parameters play important roles on the predicting protein folding rate.