Background:The aim of the present study is to analyze 215 polymerase proteins from different influenza A viruses in order to gain insight into their mutation features.Material/Methods:Using our computational mutation models to calculate differences between actual and predicted frequencies, the amino-acid pairs in a protein can be classified into randomly predictable or unpredictable portions. By comparing the percentages of both portions and the magnitudes of the differences among different proteins, we can determine which proteins have larger unpredictable portions and larger differences, and these proteins would be more sensitive to mutations according to our previous studies.Results:There are five features regarding polymerase acidic protein (PA), polymerase basic protein 1 (PB1), and polymerase basic protein 2 (PB2): (i) the PA proteins are more sensitive to mutations; (ii) the PA and PB1 proteins have experienced more mutations in the past; (iii) the PA and PB2 proteins have a larger potential for future mutations; (iv) the impact of mutations on polymerase proteins is to minimize the difference between the actual and predicted frequencies of amino-acid pairs, and (v) the avian PA proteins are more vulnerable to mutations than the human ones.Conclusions:This study provides probabilistic evidence on the current state, history, and future of different polymerase proteins from influenza A viruses towards mutation. Our analyses also highlight the mutation tendency over time and the species susceptibility.

Keywords: Influenza A virus - genetics, Mutation, Viral Proteins - genetics