Fractal Related Methods for Predicting Protein Structure Classes and Functions

The prediction of protein structure and function from amino acid sequences is one of the most important and challenge problems in molecular biology. The fractal methods and related wavelet analysis, recurrence quantification analysis (RQA) and empirical mode decomposition (EMD) methods were used to study the prediction of protein structure classes and functions. This chapter reviews the methods and results in these studies. Some of these methods include: chaos game representation of proteins and related time series; wavelet spectrum; and empirical mode decomposition and similarity of proteins. From the overall analyses, the authors found that the measure representation based on the detailed HP model and six-letter model, time series representation based on physicochemical features of amino acids, Z-curve representation, the chaos game representation of proteins can provide much information for predicting structure classes and functions of proteins. Fractal methods are useful to analyze protein sequences.