Background.Quaternary structures of proteins are closely relevant to gene regulation, signal transduction, andmany other biological\nfunctions of proteins. In the current study, a new method based on protein-conservedmotif composition in block format for feature\nextraction is proposed, which is termed block composition. Results. The protein quaternary assembly states prediction system\nwhich combines blocks with functional domain composition, called QuaBingo, is constructed by three layers of classifiers that can\ncategorize quaternary structural attributes of monomer, homooligomer, and heterooligomer.The building of the first layer classifier\nuses support vector machines (SVM) based on blocks and functional domains of proteins, and the second layer SVM was utilized\nto process the outputs of the first layer. Finally, the result is determined by the Random Forest of the third layer.We compared the\neffectiveness of the combination of block composition, functional domain composition, and pseudoamino acid composition of the\nmodel. In the 11 kinds of functional protein families, QuaBingo is 23% ofMatthews Correlation Coefficient (MCC) higher than the\nexisting prediction system. The results also revealed the biological characterization of the top five block compositions. Conclusions.\nQuaBingo provides better predictive ability for predicting the quaternary structural attributes of proteins.
Loading....