| Bioinformatics is the most effective method which analysis processing and extractuseful information of amino acid sequences. Potassium channel, which play a significant rolein physiological functions, have various of types and numbers in ion channel family, at thesame time, potassium channel is becoming a hot topic in bimolecular fields. Biologists havestudied the ion channels on its protein structures for a long time, and the amino-acidsequences are the bases factor of protein structure, so in this paper, we try to find out thedifferences between protein families based on the amino-acid sequences, and predict proteinfamilies by using different characteristic parameters. The voltage-gated potassium channelsfamilies are predicted by using bioinformatics methods and the diversity increment algorithm,which evaluated by test methods.This paper contain three part of sections. Firstly, creating voltage-gated potassiumchannel database which is propitious to our studies. Secondly, extract algorithms to classifyand predict the voltage-gated potassium channels. Based on the whole amino acid sequences,proteins families are predicted by using of the increment of diversity, with the short rangecorrelations and long range correlations and amino acid simplification information taken asthe parameter. Jack-knife test and self-consistency test methods are used for checking, whichare evaluated by three evaluating indices(sensitivity, specificity, and total accuracy). Thepredictive results show that the average prediction accuracy is better by using the amino aciddipeptide composition, and the average prediction accuracy is 99.12%. Thirdly, clusteranalysis. Based on the voltage-gated potassium channel database, only one representativesequence is selected in each subfamily, and the increment of diversity between sequencesconsist of distance matrix. Then the matrix is converted into clustering tree form, using thePHYLIP software. The results of the cluster are proved to be pretty good. |
PDF Full Text Request