Research on method of protein contact prediction of beta structure by extend Markov logic network
基于扩展的Markov逻辑网的蛋白质beta结构关联预测方法研究

In order to overcome the independent and identically distributed assumption and take advantage of rules and constrains in the protein,this paper presented a Markov logic network method to predict the beta-residue contact.This method could take advantage of Markov statistical relational learning framework to achieve the prediction of protein beta-residue contact.Experimental results show that this method can get better performance than other methods previously reported in the protein beta-residue contact prediction.It can be increased by 8% compared with BetaPro and 14 times of random prediction accuracy.