基于混合并行遗传算法和阈值限定法的基因调控网络构建

发布时间：2018-08-25 19:24

【摘要】：为了解决传统基因调控网络构建算法准确度不高的问题,提出了一种基于混合并行遗传算法和阈值限定法的新型基因调控网络构建算法。该算法分缩小解空间和参数拟合两部分,缩小解空间阶段先用奇异值分解法限定数学上可行的基因调控网络,减少不必要计算,然后用阈值限定法将每个基因的控制基因限定到一定规模,提高计算效率的同时更合乎生物信息学规则。参数拟合部分先用并行遗传算法在整个解空间快速寻优,而后采用爬山法进行小范围细致求解,提高计算精度。实验部分将本文算法应用于人类复杂疾病的皮肤黑色素瘤和2型糖尿病基因调控网络的构建上。本文计算结果与真实网络作对比,验证了本文算法的有效性。同时将本文计算结果与传统遗传算法,粒子群算法进行比较,证明本文算法具有更高的执行效率。
[Abstract]:In order to solve the problem that the accuracy of the traditional gene regulation network construction algorithm is not high, a novel gene regulation network construction algorithm based on hybrid parallel genetic algorithm and threshold limit method is proposed. The algorithm is divided into two parts: reducing solution space and parameter fitting. The singular value decomposition (SVD) method is first used to define the mathematically feasible gene control network and to reduce unnecessary calculation. Then the control gene of each gene is limited to a certain scale by the threshold restriction method, which can improve the computational efficiency and accord with the rules of bioinformatics at the same time. In the part of parameter fitting, the parallel genetic algorithm is used to quickly search the optimization in the whole solution space, and then the mountain-climbing method is used to solve the problem in a small range to improve the calculation accuracy. In the experimental part, the algorithm is applied to the construction of human skin melanoma and type 2 diabetes gene regulatory network. The results of this paper are compared with the real network, and the validity of the proposed algorithm is verified. At the same time, compared with the traditional genetic algorithm and particle swarm optimization algorithm, it is proved that the proposed algorithm is more efficient.
【作者单位】：吉林大学数学学院;梧州学院信息与电子工程学院;吉林大学计算机科学与技术学院;
【基金】：国家自然科学基金项目(61502343;61373051;61175023) 中国博士后科学基金项目(2016M590260) 广西自然科学基金项目(2015GXNSFBA139262) 梧州学院广西高校行业软件技术重点实验室项目;梧州学院院级项目(2014A002) 广西高校科研项目(KY2015ZD122) 吉林省科技发展项目(20140204004GX) 吉林大学“985工程”项目
【分类号】：Q811.4;TP18

【相似文献】