N连接糖基化过程的动态图建模

Dynamic graph modelling for N-linked glycosylation

单克隆抗体药物是当前国际市场上最具价值的药品种类之一,而N连接糖基化反应是单克隆抗体制备过程中非常重要的一个环节.本文采用有向图来描述糖蛋白结构,图模型描述反应网络,因此可以方便地采用分支限界法来生成整个反应网络.而且在这种图模型表示的反应网络中,糖基化反应的分析计算会变得更为容易.综合考虑实际细胞动态变化中高尔基体内转运反应的影响,提出了N连接糖基化反应的动态平衡方程,相比较现有的稳态模型可以提供更多的反应过程信息,并且为反应质量的实时控制做出了理论依据.另外,本文以理论计算与实测值偏差为目标,采用遗传算法对反应过程的大量生化参数进行优化,从而提高计算模型的精度.为了提升仿真模型的计算能力和对实验设计的指导作用,本文基于C++开发了一款专用的糖基化仿真计算平台,提供网络结构的可视化、反应过程的动态曲线、反应参数优化等功能.通过对比本文模型的数据分析与实际机理分析一致验证了模型的有效性.

Monoclonal antibody(m Ab) drug is one of the most valuable types of drugs currently in the international market, and N-linked glycosylation reaction is considered as an important part of the preparation of monoclonal antibody drugs. An abstract graph modeling method is proposed to describe the entire glycosylation reaction network. This graph expression can simplify the analysis and calculation of glycosylation reaction. The dynamic mass balance equations of N-linked glycosylation reaction are also discussed which can provide more information than steady-state model and theoretical basis for real-time control of drug quality. In order to facilitate the simulation of dynamic model and to guide the experimental design, we develop a dedicated glycosylation simulation platform based on C++. Simulation platform provides dynamic calculation of glycan concentration, visualization of network structure and optimization of reaction parameters based on genetic algorithm. At last, the glycosylation reaction model is verified based on the simulation platform whose results are consistent to experiment and mechanism analysis.