随机布尔网络吸引子的鲁棒性与相对稳定性
发布时间：2018-11-28 20:46
【摘要】：生物网络中时常会出现多个表型或多条动力学路径共存的现象,但其表型和动力学路径在随机涨落影响下的稳定性问题迄今尚未得到完整的认识.本文通过分析随机布尔网络模型来尝试回答这个问题.表型和动力学路径的稳定性对应于随机布尔网络吸引子的鲁棒性和相对稳定性,而后者可以用指数扰动马氏链理论来加以刻画和分析.已有的指数扰动马氏链理论已经告诉我们,从某个吸引子的吸引域溢出的时间的对数正比于该吸引域的"非平衡态活化能势垒",于是本文首先推广了这个理论,证明了指数扰动马氏链模型中吸引子之间的最佳转移路径都是等概率的,因此在吸引子间"非平衡态活化能势垒"相等的情形下,吸引子的相对稳定性是由最佳转移路径的个数来决定的.该理论还预示着,当模型中的随机涨落很小时,将会出现如下的相变现象:在参数空间的某一区域内,这些表型和动力学路径将以某一比例共存在;而在另一些区域内,某个表型或某条动力学路径将占据主导地位,从而成为全局吸引子.最后,在人造模型以及蛋白质p53动力学模型中应用和验证了该理论,并且通过计算吸引子稳定性对于具体动力学参数的敏感性,提供了一种辨别网络中重要节点和节点间重要相互作用的新方法.
[Abstract]:There are many phenotypes or multiple dynamic paths coexisting in biological networks, but the stability of their phenotypes and kinetic pathways under the influence of random fluctuations has not been fully understood. This paper attempts to answer this question by analyzing the stochastic Boolean network model. The stability of phenotypic and dynamic paths corresponds to the robustness and relative stability of attractors in stochastic Boolean networks, which can be characterized and analyzed by exponential perturbation Markov chain theory. The existing exponential perturbed Markov chain theory has told us that the logarithm of the spillover time from an attractor is proportional to the "non-equilibrium activation energy barrier" of the attraction domain. It is proved that the optimal transfer paths between attractors in the exponential perturbed Markov chain model are equiprobability, so in the case of equal "non-equilibrium activation energy barrier" among attractors, The relative stability of attractors is determined by the number of optimal transition paths. The theory also indicates that when the random fluctuations in the model are very small, the following phenomena will occur: in a certain region of the parameter space, these phenotypic and kinetic pathways will co-exist in a certain proportion; In other regions, a phenotype or a dynamic path will dominate and become a global attractor. Finally, the theory is applied and validated in artificial model and protein p53 kinetic model, and the sensitivity of attractor stability to specific kinetic parameters is calculated. A new method for identifying important interactions between important nodes and nodes in the network is presented.
【作者单位】： Department
【基金】：国家自然科学基金(批准号:10901040,21373021和11622101) 中国优秀博士论文作者基金会(批准号:201119)资助项目
【分类号】：O211
，
本文编号：2364214
[Abstract]:There are many phenotypes or multiple dynamic paths coexisting in biological networks, but the stability of their phenotypes and kinetic pathways under the influence of random fluctuations has not been fully understood. This paper attempts to answer this question by analyzing the stochastic Boolean network model. The stability of phenotypic and dynamic paths corresponds to the robustness and relative stability of attractors in stochastic Boolean networks, which can be characterized and analyzed by exponential perturbation Markov chain theory. The existing exponential perturbed Markov chain theory has told us that the logarithm of the spillover time from an attractor is proportional to the "non-equilibrium activation energy barrier" of the attraction domain. It is proved that the optimal transfer paths between attractors in the exponential perturbed Markov chain model are equiprobability, so in the case of equal "non-equilibrium activation energy barrier" among attractors, The relative stability of attractors is determined by the number of optimal transition paths. The theory also indicates that when the random fluctuations in the model are very small, the following phenomena will occur: in a certain region of the parameter space, these phenotypic and kinetic pathways will co-exist in a certain proportion; In other regions, a phenotype or a dynamic path will dominate and become a global attractor. Finally, the theory is applied and validated in artificial model and protein p53 kinetic model, and the sensitivity of attractor stability to specific kinetic parameters is calculated. A new method for identifying important interactions between important nodes and nodes in the network is presented.
【作者单位】： Department
【基金】：国家自然科学基金(批准号:10901040,21373021和11622101) 中国优秀博士论文作者基金会(批准号:201119)资助项目
【分类号】：O211
，
本文编号：2364214
本文链接：https://www.wllwen.com/kejilunwen/yysx/2364214.html
