Poster Presentation 2014 International Biophysics Congress

Antifreeze protein is a kind of special functional protein that exists in the organisms which live in cold environment. Antifreeze proteins can bind onto the surface of the ice crystals. This binding causes a non-colligative depression of the freezing point. The difference between melting point and freezing point of ice crystals is termed thermal hysteresis [1]. Therefore, AFPs is also called thermal hysteresis protein. The binding of AFPs to ice also inhibits the natural recrystallization of ice when numerous ice crystals are present. Choosing a set of reasonable information parameters from protein sequence is very helpful for predicting the antifreeze protein. The hydropathy and the chemical structure of the R group of amino acid are important physicochemical characteristics for identifying AFPs. Based on the n-Peptide compositions and these physicochemical characteristics, an algorithm of Support Vector Machine (SVM) [2] is proposed for predicting antifreeze proteins. AFPs are also predicted with high level of accuracy, from their primary amino acid sequence. It is also evidence that the primary sequence contains important information which determines protein advance structure. Through the analysis of AFPs amino acid sequence motifs by MEME, we can further identify the key functional residues of the ice-binding surfaces. For the same physicochemical characteristics, the predictive results of dipeptide compositions are relatively higher than that obtained by amino acid composition.