基于GES机制遗传算法构建基因调控网络

目的：基因调控网络在药物研发与疾病防治方面有重要的生物学意义。目前基于芯片数据构建网络的方法普遍效率不高，准确度较低，为此提出了一种新的高效调控网络结构预测算法。方法：提出了一种基于贪婪等价搜索机制的遗传算法构建基因调控网络模型。通过引入遗传算法的多点并行性，使得算法易于摆脱局部最优。通过编码网络结构作为遗传算法的染色体和设计基于GES机制的变异算子,使网络的进化过程基于马尔科夫等价空间而不是有向无环图空间。结果：通过对标准网络ASIA和酵母调控网络的预测，与近期Xue-wen Chen等提出的Order K2算法进行了比较，在网络构建准确率上获得了更佳的结果。与标准遗传算法比较下在执行效率上大大提高。结论：提出的算法在网络结构预测准确率上相对于最近提出的Order K2算法在准确率上效果更佳，并且相较标准遗传算法网络在进化过程上效率更高。

Construction of Gene Regulatory Networks Based on Genetic Algorithm of GES Mechanism

Motivation： Gene regulatory network has important biological meaning in development of medicine and prevention of diseases. Current methods of construction of GRN based on microarray data generally are lack of efficiency and accuracy. Method： In this paper, a novel method is presented to predict gene regulatory networks based on Genetic Algorithm of Greedy Equivalence Search Mechanism. By means of coding the network structure as the chromosome of GA and designing the mutation operators based on GES mechanism, the process of network evolve in Markov Equivalence space, rather than in DAG （Directed Acyclic Graph： DAG） space. Result： With comparing our algorithm to Order K2 developed by Xue-wen Chen et al in predicting the structure of Asia and yeast genes’ regulatory network, our algorithm got more accurate result, and converged more fast than canonical GA. Conclusion： GA of GES mechanism is more accurate in comparision to Order K2, the method recently developed by Chen et al. by introduction of GES mechanism, our method is more efficient than canonical GA.