【作者】 何鹏；

【导师】 周真；

【作者基本信息】 哈尔滨理工大学 ， 精密仪器及机械， 2013， 博士

【摘要】 随着检测理论和技术的不断发展，检测的对象已经从传统的检测领域延伸到图像、化学、生物、医药等领域，在生物制药、食品、化工等领域，蛋白质聚集现象是普遍存在的。对它动态的聚集过程进行检测一直是学者们研究的热点之一。对他的研究将对生物药中蛋白质聚集导致药效的降低、是否引起免疫应答、相关的毒理学效应、药的保质期等研究都有着重要的意义。目前蛋白质的聚集过程主要依赖于经验公式、质量作用定律、相关的数学模型来预测其聚集结果。若能在生产中不断地控制初始条件，加上对中间结果的检测来修正初始条件，这样的闭环控制就能尽最大可能防止蛋白质的聚集。如何控制初始条件？聚集过程中哪个时间点更能有效地来修正初始条件？都要依赖于聚集过程的动力学方程。目前的聚集动力学方程的模型都是假定：1蛋白质是某一形貌，如：球状、杆状等简单的几何形状；2在假定某种形貌的基础上，计算聚集体的尺寸。3聚集过程是个动态的，它包含成核、聚集、生长和解聚行为，在前两个基础上，再对模型参数进行识别，即获得方程的初始条件。这无疑缺少普适性，蛋白质的形貌是复杂的，甚至是变化的。本文采用了分形维来描述和表达聚集体的形貌，而不是假定某一几何形貌，以此建立了新的聚集体动力学模型，这一模型可以同时反映出聚集体的尺寸和分形维两个特征。增加了使用的普适性。同时对模型的初始条件，即模型参数进行了参数估计和辨识。并用仿真和实验进行了验证工作。并给出了较满意的结果。本文的主要工作可归纳如下：1.采用分形理论和群体粒子衡算方程(Population Balance Equation,PBE)，提出了描述蛋白质聚集过程的动力学模型，推导了动力学模型的增长函数和成核函数。并通过计算机仿真实现了动态尺寸模拟、动态分形维模拟。该模型可以同时给出蛋白质聚集体的尺寸和分形维两个关键参数的动态尺寸。针对聚集体尺寸和分形维的计算中存在的问题，应用有限元盖拉金法编制了计算机软件求解了动力学方程。2.在闭环控制中，要不停修正初始条件即模型参数。在模型参数的辨识过程中，提出了一种新的基于K-L信息距离的蛋白质聚集过程模型参数辨识方法。通过优化算法，给出了蛋白质聚集过程的初始条件和实验检测时间点，该方法将指导实验验证工作中的实验设计，并选择了具有普适性意义溶菌酶的聚集实验。3.将上述方法应用于溶菌酶的聚集过程控制，利用参数辨识给出的初始条件、时间检测点，分别应用激光粒度仪和密度计测定溶菌酶冷却聚集过程中尺寸和浓度变化的一般规律。根据这些变化规律，精确拟合聚集动力学参数的基于尺寸和浓度变化规律。实验证明，采用分形理论和PBE方程提出描述蛋白质聚集过程的仿真动力学模型可以有效预测聚集体形貌及其分布。4.针对聚集过程中难以测量存在的宏观聚集现象，研究了大尺寸分形维测量问题。以不同稀释条件的添加凝乳酶牛奶溶液为实验材料，提出了一种基于PBE模拟模型的致密聚集体分形维测量新方法，方法在不建立聚集体互相关函数的情况下，测量了蛋白质聚集体大尺寸分形维。5.建立蛋白质聚集体过程测量系统，该系统验证了论文中所提出的的观点，利用所提出的模型，研究了鸡蛋清溶菌酶间歇冷却聚集过程中，温度条件、pH值条件对聚集过程的影响。实验测量发现溶菌酶在不同温度、pH值条件下，聚集过程差异较大，而提出的模型能很好地预测这一点。进一步研究了聚集体光散射特性和分形特性、聚集体尺寸分布、分形维分布和聚集体粒子数之间的关系，为研究聚集体聚集过程检测仪器奠定了基础。更多还原

【Abstract】 With the continuous development of detection theory and technology,detection object has extended to image, chemistry, biology, medicine and otherfields from the traditional detection field, protein aggregation phenomenon iswidespread in the field of biological pharmacy, food, chemical and so on. Studyon protein dynamic accumulation process is one of hot topics in the study ofscholars, the research is of great significance for protein accumulation resultingin a loss of efficacy, cause immune response, related toxicological effect, such asthe shelf life of medicine in biological medicine.The protein aggregation process mainly depends on experience formula, thelaw of mass action, the relevant mathematical model to predict the results. If wecan continuously control initial conditions in the production process, using thetested intermediate results to modify initial conditions, the closed loop controlcan prevent protein concentrations as much as possible. How to control the initialcondition and which time points in the process of aggregation is more effective tomodify initial conditions, depends on dynamic equation of the accumulationprocess. At present, the models of accumulation dynamics equation are allassumed:1protein belongs to a particular morphology, such as spherical, rodsand other simple geometric shapes;2On the basis of assumption that certainmorphology, calculate the size of aggregate particles;3Accumulation is adynamic process and it includes nucleation, aggregation, growth anddepolymerization behavior, on the basis of the former two behaviors,identification of model parameters, obtain the initial conditions of equation, itcan be seen that this method is lack of universality, because of the topography ofprotein is complex and changing. In this paper, fractal dimension method is usedto describe and express the aggregation morphology, instead of assuming ageometric shape, in order to establish a new aggregate dynamic model. Thismodel can reflect aggregate size and fractal dimensions characteristics at the same time, increase the universality, initial conditions of the model, parameterestimation and identification of model parameters at the same time, simulationand experimental work is used for verify the proposed model and the satisfactoryresult is given.In this research, the main work can be summarized as follows:1. Adopt fractal theory and population balance equation (PBE), dynamicmodel is proposed for describe protein aggregation process, the growth functionand kernel function of dynamic model was deduced. The proposed modelaccomplishment dynamic size and dynamic fractal dimension simulation bycomputer simulation. The model can parse the size and fractal dimension ofprotein aggregation at the same time. In view of calculation problems ofaggregate size and fractal dimension, finite element cover Larkin method is usedto compile computer software and the purpose is to solve dynamic equations.2. In the process of closed loop control, constantly revised initial conditions,that is model parameters. Aim at the identification problem of aggregationdynamic model, a kind of protein aggregation process model parameteridentification method is proposed based on K-L information distance, optimizingthe algorithm, initial conditions and experimental testing point of proteinaggregation process were given, this method can guide the experimental designof experimental verification work and select lysozyme aggregate experiment ofuniversal significance.3. This method was applied to lysozyme aggregation process control, makeuse of the given initial conditions, time check points by parameter identificationmethod, using a laser granulometer and density gauge measuring general rule ofsize and concentration change during lysozyme cooling gathered process,according to the change rule, precise fitting size and concentration change rulebased on aggregation kinetics parameters. Experiments show that dynamic modelof describe protein aggregation process based on fractal theory and populationbalance equation (PBE) is proposed, the model can effectively predict theaggregation morphology and distribution.4. Aimed at the Macro aggregation phenomenon in the process ofaggregation is difficult to measured, large size particles fractal dimensionmeasurement problem is studied, with different dilution condition to add rennetin milk solution as the experimental materials, a new dense aggregates fractal dimension measuring method based on PBE simulation model is proposed, in thecase of don’t build aggregation cross-correlation function, measuring proteinaggregation large size fractal dimension.5. Establish a protein aggregation process measurement system, the systemis used to verify the effectiveness of the proposed method in the dissertation, inthe process of hen egg white lysozyme intermittent cooling condition, theinfluence of temperature, pH value and other condition on the accumulationprocess is studied by the proposed model. The results showed that lysozymeaccumulation process difference is bigger under the condition of differenttemperature and pH value, the PBE mathematical model has a good predictionwith accumulation process difference. Systematically studied the relationshipbetween protein aggregation light scattering characteristics and fractal features、aggregation size distribution、 fractal dimension distribution and proteinaggregate particles, the article proposed method laid a foundation for the study ofaggregation process instrumentation.更多还原