Research Of Protein Remote Homology Detection Based On Pseudo Amino Acids

How to design a precise and fast method for predicting protein remote homology is a problem that has been puzzling researchers for a long time in the field of bioinformatics.The main task of protein remote homology detection is to find protein sequences related to query proteins with evolutionary information in a protein database of known structure and function.Researchers have put forward many prediction methods based on statistical computing,which can be roughly divided into three categories: 1)methods based on comparison;2)methods based on discriminant;and 3)methods based on ranking.However,these methods are computationally intensive,inefficient,and not so ideal for the detection of proteins with very low sequence similarity.In the rapid development of the information age,unknown protein sequences are gradually increasing and the sequence similarity between most important homologous proteins is low.How to find an efficient detection method under such circumstances is still a problem need to be overcome by researchers and bioinformatician.In this paper,based on the existing research results,we adopt the discriminant method(binary classification)to study the existing data sets.A variety of feature extraction methods were used for protein sequences,such as amino acid composition(ACC),dipeptide composition(DP),correlation factors derived from amino acid physicochemical properties(SOF),Position Specific Scoring Matrix(PSSM)and the Grey PSSM model(Grey-PSSM).All the proposed features are input into the classifier C-SVC in the LIBSVM package for prediction.Secondly,we also use the ranking method based on the features from Grey-PSSM and Cellular Automaton Image(PCA-GLCM).Through the comparison of the results obtained by a series of feature extraction and classification methods,an effective method for detection of remote homologous proteins is designed.It is hoped that it will help and promote the study of remote homologous protein detection in the future.