WO2023019883A1 - 利用细胞代谢网络监测生物制造过程的方法 - Google Patents

利用细胞代谢网络监测生物制造过程的方法 Download PDF

Info

Publication number
WO2023019883A1
WO2023019883A1 PCT/CN2022/073490 CN2022073490W WO2023019883A1 WO 2023019883 A1 WO2023019883 A1 WO 2023019883A1 CN 2022073490 W CN2022073490 W CN 2022073490W WO 2023019883 A1 WO2023019883 A1 WO 2023019883A1
Authority
WO
WIPO (PCT)
Prior art keywords
monitoring
metabolic
flux
cell
metabolic flux
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2022/073490
Other languages
English (en)
French (fr)
Inventor
刘飞
李�权
赵忠盖
李恭新
王志国
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangnan University
Original Assignee
Jiangnan University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Jiangnan University filed Critical Jiangnan University
Publication of WO2023019883A1 publication Critical patent/WO2023019883A1/zh
Priority to US18/396,157 priority Critical patent/US20240273262A1/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B5/00—ICT specially adapted for modelling or simulations in systems biology, e.g. gene-regulatory networks, protein interaction networks or metabolic networks
    • G16B5/30—Dynamic-time models
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
    • G16B40/20—Supervised data analysis
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B50/00—ICT programming tools or database systems specially adapted for bioinformatics
    • G16B50/30—Data warehousing; Computing architectures

Definitions

  • the invention relates to the technical field of industrial bioprocess control, in particular to a method for monitoring a biomanufacturing process using a cell metabolism network.
  • biomanufacturing process operation status assessment and fault monitoring have important practical engineering value.
  • the technical problem to be solved by the present invention is to overcome the problem that the microscopic metabolic activities in the cells cannot be observed in the prior art, resulting in insufficient evaluation of the existing fermentation production operation status, and to provide a method for monitoring the biomanufacturing process using the cell metabolic network.
  • This method not only utilizes the detectable or observable information in the microbial manufacturing process, but also fully mines the undetectable intracellular metabolic information in the process of cell growth and reproduction, which can more effectively implement the evaluation of the operating status and fault monitoring of the manufacturing process.
  • the present invention provides a method for monitoring the biomanufacturing process by using the cell metabolic network, comprising the following steps: performing flux balance analysis based on the metabolic network to obtain metabolic flux reflecting the process of cell growth and reproduction;
  • the internal metabolic flux and the production process variables constitute the training data set, and the monitoring model and statistical control limits of the biological fermentation process are established; based on the monitoring model, the real-time collected production process variables and metabolic flux calculation values are used for online monitoring of the operating status.
  • the flux balance analysis process includes: establishing a flux balance equation; establishing a dynamic relationship between the concentration of extracellular metabolites and metabolic flux; metabolic flux.
  • the metabolic flux changes with time, and the metabolic flux change law between two adjacent time points is collected to describe the transient characteristics of the metabolic flux, including constants, linear functions, Quadratic function.
  • multi-batch extracellular metabolite concentration sampling values are selected according to actual production, one or more samples are taken within the calculation time interval, and then curve fitting is performed.
  • the optimal metabolic flux is obtained through optimization algorithms of derivative, linear programming, and quadratic programming.
  • establishing a monitoring model of a biological fermentation process includes: combining production process variables and intracellular metabolic fluxes to form an extended sample data set; standardizing the sample data set to generate a standard sample data set; using the principal element Analyze and establish monitoring models; determine control limits for monitoring statistical indicators.
  • the intracellular metabolic flux is selected from all metabolic fluxes, intracellular and extracellular exchange metabolic fluxes, or internal metabolic fluxes.
  • the principal component analysis includes general principal component analysis, multi-stage principal component analysis, kernel principal component analysis and support vector machine methods.
  • the online monitoring of the operating status includes: collecting the concentration of extracellular metabolites and production process variables of the current batch at the current moment; the metabolic flux obtained by estimating the concentration of extracellular metabolites; Variables and metabolic fluxes are combined to form a test data set; the test data set is standardized to generate a test standard data set; the test data to be filled in the future is estimated; the statistical standard is calculated and compared with the statistical control limit.
  • the method of estimating and filling the test data at the future time includes directly using the mean data of the corresponding time of the previous batches, or considering the difference between the current and past actual data and the corresponding mean data of this batch, Estimates are populated with future data.
  • the method for monitoring the biomanufacturing process using the cell metabolic network described in the present invention uses the cell metabolic network and the concentration of extracellular observable metabolites on the basis of multi-batch production process data to calculate the metabolic flux by estimating the cellular metabolic flux Combined with the variables of the production process, an extended sample data set is constructed, and a method for online monitoring of the process operation status is given.
  • This method not only utilizes the detectable or observable information of the microbial manufacturing process, but also fully mines the undetectable cell growth and reproduction process. Internal metabolism information can more effectively implement the evaluation of the operation status of the manufacturing process and fault monitoring, and can detect abnormalities in the process operation in time so that corresponding control measures can be taken, which has a significant effect on reducing safety hazards and improving the overall benefits of the enterprise .
  • Fig. 1 is the step diagram of the method for monitoring the biomanufacturing process by using the cell metabolic network in the present invention
  • Fig. 2 is a schematic flow chart of a method for monitoring a biomanufacturing process using a cell metabolic network in the present invention
  • Fig. 3 Calculation results of the concentration of observable metabolites in the fermentation process of penicillin.
  • Fig. 5 Monitoring effect of the operation status of the penicillin fermentation process.
  • the method for monitoring the biomanufacturing process using a cell metabolic network of the present invention includes the following steps: performing flux balance analysis based on the metabolic network to obtain a metabolic flux reflecting the process of cell growth and reproduction; Quantities and production process variables together constitute a training data set, and establish a monitoring model and statistical control limits for biological fermentation processes; based on the monitoring model, use real-time collected production process variables and metabolic flux calculation values for online monitoring of operating conditions; Based on the production process data, using the cell metabolic network and the concentration of extracellular observable metabolites, by estimating the cell metabolic flux, combining the metabolic flux with the production process variables, constructing an extended sample data set, and giving an online monitoring process operation
  • the status method which not only utilizes the detectable or observable information in the microbial manufacturing process, but also fully mines the undetectable intracellular metabolic information in the process of cell growth and reproduction, can more effectively implement the evaluation of the operating status of the manufacturing process and fault monitoring, Being able to detect abnormalities in the process operation in time so that corresponding
  • the flux balance analysis process includes: establishing a flux balance equation; establishing the dynamic relationship between the concentration of extracellular metabolites and metabolic flux; estimating the metabolic flux according to the sampling values of the concentration of extracellular metabolites ;
  • Establishing a monitoring model for biological fermentation process includes: combining production process variables and intracellular metabolic fluxes to form an extended sample data set; standardizing the sample data set to generate a standard sample data set; using principal component analysis to establish a monitoring model; determining monitoring statistics
  • the control limit of the index; the online monitoring of the operation status includes: collecting the concentration of extracellular metabolites and production process variables at the current moment of the current batch; the metabolic flux obtained by estimating the concentration of extracellular metabolites; combining the production process variables with the metabolic flux Quantities are combined to form a test data set; the test data set is standardized to generate a test standard data set; the test data to be filled in the future is estimated; the statistical standard is calculated and compared with the statistical control limit.
  • the net accumulation or consumption rate of intracellular metabolites in the microbial fermentation process is much smaller than the exchange reaction rate between extracellular and extracellular metabolites.
  • the number is F and contains m intracellular metabolites and R extracellular metabolites.
  • the flux balance equation is as follows:
  • C R , t ⁇ R R represents the concentration vector of R measurable extracellular metabolites at time t
  • S R ⁇ R R ⁇ F represents the stoichiometric matrix of extracellular measurable metabolites
  • V t ⁇ R F represents Metabolic flux vector in the metabolic network, including intracellular metabolic flux and intracellular exchange flux
  • S m ⁇ R m ⁇ F represents the stoichiometric matrix of m intracellular metabolites.
  • u D represents the free flux at the Dth calculation time
  • k(t, t D ) is the coefficient function of the linear change of metabolic flux in each time interval, which is selected according to Table 1.
  • the calculated value of the concentration at time t′ n is:
  • C r, 0 represents the initial concentration value of the rth extracellular metabolite.
  • the sampling point can be selected according to the actual production.
  • the first and last calculation time points of the metabolic flux V t should correspond to the initial and final moments of actual sampling observations, respectively.
  • the initial concentration C R0 and the calculation time t D of metabolic flux can be added as optimization variables, and the optimization algorithm can be used to solve Obtain the optimal free flux U, initial concentration C 0 and time t D .
  • the extended sample data set is standardized and transformed into a standard training data set with a mean of 0 and a variance of 1, which approximately obeys a multidimensional normal distribution:
  • mean(G) is the mean value of the same variable of each batch in G at the same time, that is, the average running trajectory of multiple batches under normal operation
  • std(G) is the standard deviation of the same variable in each batch of G at the same time.
  • the monitoring model can also adopt methods such as multi-stage multi-directional principal component analysis, multi-directional kernel principal component analysis and multi-class support vector machine.
  • the data from n+1 after the current time to the end of the batch N is filled with the estimated method, and filled with the standardized G test, the data value of the nth time in n, to obtain G′ test, n ⁇ R N ⁇ H , and then expanded into G′′ test in the time direction, n ⁇ R 1 ⁇ N H .
  • ⁇ RA ⁇ A is a diagonal matrix composed of elements in the ith batch score vector of G train .
  • T 2 and SPE n are lower than the statistical control limit, the process is running normally; otherwise, there is an abnormality or failure in the early warning process.
  • the metabolic network of penicillin fermentation contains 66 metabolic fluxes (including 5 intracellular exchange fluxes) and 49 intracellular metabolites, and 5 extracellular metabolites, including glucose, penicillin, bacterial biomass, oxygen and carbon dioxide, can be detected , considering a total of 11 production process variables such as reaction temperature, pH value, ventilation rate, stirring power, substrate flow acceleration rate and cooling water flow acceleration rate, and intracellular and extracellular exchange metabolic fluxes, 30 batches with 800 sampling points per batch were selected. Form an extended sample data set.
  • Figure 3 shows the estimated value of intracellular and extracellular exchange metabolic flux, thus expanding the sample data set not only includes biological fermentation
  • the process-detectable production process variable information also includes the unmeasurable metabolic information inside the cell, which more comprehensively reflects the characteristics of the production process.
  • Figure 4 shows the online monitoring results of the present invention:
  • the process monitoring represented by (a) only uses the production process variables, while (b) combines the metabolic network information at the same time. It can be found from the T 2 monitoring diagram that when a fault occurs, the method proposed by the present invention can monitor the fault in time and accurately, but there is a certain time lag in monitoring based on production variables.
  • the embodiments of the present application may be provided as methods, systems, or computer program products. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
  • computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
  • These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions
  • the device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Artificial Intelligence (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • General Health & Medical Sciences (AREA)
  • Biotechnology (AREA)
  • Biophysics (AREA)
  • Software Systems (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Bioethics (AREA)
  • Epidemiology (AREA)
  • Public Health (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Molecular Biology (AREA)
  • Physiology (AREA)
  • Apparatus Associated With Microorganisms And Enzymes (AREA)

Abstract

本发明涉及一种利用细胞代谢网络监测生物制造过程的方法，包括如下步骤：基于代谢网络进行通量均衡分析，得到反映细胞生长与繁殖过程的代谢通量；将细胞内代谢通量与生产过程变量一起构成训练数据集，建立生物发酵过程监测模型及统计量控制限；基于监测模型，利用实时采集的生产过程变量和代谢通量计算值进行运行状况在线监测。本发明的利用细胞代谢网络监测生物制造过程的方法，该方法既利用微生物制造过程可检测或可观测信息，又充分挖掘细胞生长繁殖过程中不可检测的胞内代谢信息，能够更加有效的实施对生产制造过程运行状况评估及故障监测。

Description

利用细胞代谢网络监测生物制造过程的方法 技术领域
本发明涉及工业生物过程控制技术领域，尤其是指一种利用细胞代谢网络监测生物制造过程的方法。
背景技术
随着全球经济发展和社会进步，为了满足人们日益增长的生活需求，以工业微生物技术为代表的小批量、高利润、多品种生产方式受到重视。生物制造过程一般在生物反应器中进行，其可以通过批次生产的方式在一定周期时间内获得经济效益高的产物，比如常用抗生素青霉素的生产，其外源分泌蛋白等一般由生物发酵反应而得。然而在发酵生产过程中经常会出现各类生产状况的失控或运行故障，如果不能及早监测、预警并及时调控，最终会影响生产质量和目标产物的产量。因此，生物制造过程运行状况评估与故障监测，具有重要的实际工程价值。
生物发酵过程中同时存在物化和生化反应，细胞内部机理复杂，而实际生产过程中一般只能检测到温度、pH、通风率、搅拌功率等基本生产工艺变量及一些可观测的细胞外物质浓度，对于细胞内的微观的代谢活动则无法观测，导致现有的发酵生产运行状况评估技术主要依据生产过程中各工艺变量和物质浓度进行综合分析，其监测的实效性和及时性有待提高。
发明内容
为此，本发明所要解决的技术问题在于克服现有技术中于细胞内微观的代谢活动无法观测，导致现有的发酵生产运行状况评估不足的问题，提供一种利用细胞代谢网络监测生物制造过程的方法，该方法既利用微生物制造过程可检测或可观测信息，又充分挖掘细胞生长繁殖过程中不可检测的胞内代 谢信息，能够更加有效的实施对生产制造过程运行状况评估及故障监测。
为解决上述技术问题，本发明提供了一种利用细胞代谢网络监测生物制造过程的方法，包括如下步骤：基于代谢网络进行通量均衡分析，得到反映细胞生长与繁殖过程的代谢通量；将细胞内代谢通量与生产过程变量一起构成训练数据集，建立生物发酵过程监测模型及统计量控制限；基于监测模型，利用实时采集的生产过程变量和代谢通量计算值进行运行状况在线监测。
在本发明的一个实施例中，所述通量均衡分析过程包括：建立通量均衡方程；建立胞外代谢物浓度与代谢通量的动态关系；根据多批次胞外代谢物浓度采样值估计代谢通量。
在本发明的一个实施例中，所述代谢通量的随时间变化，采集相邻两个时间点之间的代谢通量变化规律对代谢通量进行瞬态特征描述，包括常数、线性函数、二次函数。
在本发明的一个实施例中，多批次胞外代谢物浓度采样值根据生产实际来选取，在计算时间间隔内采样一次或多次采样，再进行曲线拟合。
在本发明的一个实施例中，在估计代谢通量时，通过求导数、线性规划、二次规划的优化算法求得最优代谢通量。
在本发明的一个实施例中，建立生物发酵过程监测模型包括：将生产过程变量和细胞内代谢通量合并构成扩展样品数据集；对样品数据集进行标准化处理生成标准样品数据集；利用主元分析建立监测模型；确定监测统计指标的控制限。
在本发明的一个实施例中，在扩展样品数据集时，所述细胞内代谢通量选取全部代谢通量或胞内外交换代谢通量或内部代谢通量。
在本发明的一个实施例中，所述主元分析包括一般主元分析、多阶段主元分析、核主元分析以及支持向量机方法。
在本发明的一个实施例中，所述运行状况在线监测包括：采集当前批次当前时刻的胞外代谢物浓度和生产过程变量；由胞外代谢物浓度估计获得的 代谢通量；将生产过程变量与代谢通量合并构成测试数据集；对测试数据集进行标准化处理生成测试标准数据集；预估填充未来时刻的测试数据；计算统计标准，并与统计量控制限比较。
在本发明的一个实施例中，预估填充未来时刻的测试数据的方法包括直接用前若干批次对应时刻的均值数据、或考虑本批次当前及过去时刻实际数据与对应均值数据之差，预估填充未来数据。
本发明的上述技术方案相比现有技术具有以下优点：
本发明所述的利用细胞代谢网络监测生物制造过程的方法，在多批次生产过程数据基础上，利用细胞代谢网络和胞外可观测代谢物浓度，通过估计细胞代谢通量，将代谢通量与生产过程变量结合，构建扩展样本数据集，给出一种在线监测过程运行状况的方法，该方法既利用微生物制造过程可检测或可观测信息，又充分挖掘细胞生长繁殖过程中不可检测的胞内代谢信息，能够更加有效的实施对生产制造过程运行状况评估及故障监测，能够及时发现过程运行中的异常，以便做出相应的控制措施，对于减少安全隐患、提高企业综合效益具有明显的作用。
附图说明
为了使本发明的内容更容易被清楚的理解，下面根据本发明的具体实施例并结合附图，对本发明作进一步详细的说明，其中
图1是本发明利用细胞代谢网络监测生物制造过程的方法步骤图；
图2是本发明利用细胞代谢网络监测生物制造过程的方法流程示意图；
图3青霉素发酵过程部分可观测代谢物浓度计算结果。
图4青霉素发酵过程部分代谢通量动态估计。
图5青霉素发酵过程运行状况监测效果。
具体实施方式
下面结合附图和具体实施例对本发明作进一步说明，以使本领域的技术 人员可以更好地理解本发明并能予以实施，但所举实施例不作为对本发明的限定。
参照图1所示，本发明的利用细胞代谢网络监测生物制造过程的方法，包括如下步骤：基于代谢网络进行通量均衡分析，得到反映细胞生长与繁殖过程的代谢通量；将细胞内代谢通量与生产过程变量一起构成训练数据集，建立生物发酵过程监测模型及统计量控制限；基于监测模型，利用实时采集的生产过程变量和代谢通量计算值进行运行状况在线监测；在多批次生产过程数据基础上，利用细胞代谢网络和胞外可观测代谢物浓度，通过估计细胞代谢通量，将代谢通量与生产过程变量结合，构建扩展样本数据集，给出一种在线监测过程运行状况的方法，该方法既利用微生物制造过程可检测或可观测信息，又充分挖掘细胞生长繁殖过程中不可检测的胞内代谢信息，能够更加有效的实施对生产制造过程运行状况评估及故障监测，能够及时发现过程运行中的异常，以便做出相应的控制措施，对于减少安全隐患、提高企业综合效益具有明显的作用。
参照图2所示，所述通量均衡分析过程包括：建立通量均衡方程；建立胞外代谢物浓度与代谢通量的动态关系；根据多批次胞外代谢物浓度采样值估计代谢通量；建立生物发酵过程监测模型包括：将生产过程变量和细胞内代谢通量合并构成扩展样品数据集；对样品数据集进行标准化处理生成标准样品数据集；利用主元分析建立监测模型；确定监测统计指标的控制限；所述运行状况在线监测包括：采集当前批次当前时刻的胞外代谢物浓度和生产过程变量；由胞外代谢物浓度估计获得的代谢通量；将生产过程变量与代谢通量合并构成测试数据集；对测试数据集进行标准化处理生成测试标准数据集；预估填充未来时刻的测试数据；计算统计标准，并与统计量控制限比较。
建立通量均衡方程：
微生物发酵过程中胞内代谢物的净积累或消耗速率远小于胞内外代谢物之间的交换反应速率，对于一个总化学反应个数为F并且包含了m个胞内代谢物和R个胞外可测代谢物的代谢网络，其通量均衡方程如下：
Figure PCTCN2022073490-appb-000001
其中,C R，t∈R R表示R个可测的胞外代谢物在t时刻浓度向量，S R∈R R×F表示胞外可测代谢物的化学计量矩阵，V t∈R F表示代谢网络内代谢通量向量，包括胞内代谢通量及胞内外交换通量，S m∈R m×F表示m个胞内代谢物的化学计量矩阵。
微生物代谢网络中化学反应个数F多于细胞内代谢物数量m，其自由度d＝F-rank(S m)，采用一组独立通量u t∈R d的线性组合表示代谢通量，即V t＝K·u t，其中K∈R F×d表示S m零空间的一组标准正交基，u t称为自由通量。
代谢通量的瞬态特征描述：
考虑代谢通量V t随时间变化，引入计算时间序列点t D，D＝1，...，Z，用以描述分析V t的动态变化，即将整个生产过程划分为Z-1个时间间隔，在每个时间间隔内，即在相邻两个时间点t D和t D+1之间，假设代谢通量按一定函数规律变化，可以是常数、线性函数、二次型函数。选用线性函数后的代谢通量V t表示如下：
Figure PCTCN2022073490-appb-000002
其中，u D表示第D个计算时刻的自由通量，k(t，t D)是各时间间隔内代谢通量线性变化的系数函数，按表1选取。
Figure PCTCN2022073490-appb-000003
表1
基于代谢通量的胞外代谢物浓度计算：
将式(2)代入式(1)，得：
Figure PCTCN2022073490-appb-000004
对式(3)两边积分得：
Figure PCTCN2022073490-appb-000005
其中，函数
Figure PCTCN2022073490-appb-000006
而C R0＝[C 1，0，C 2，0，...，C R，0] T表示R个胞外可观测代谢物的初始浓度向量。
生产过程中实际采样观测时间点为t′ n,n＝1，...，N，根据式(4)，第r个胞外代谢物，r＝1，...，R，在实际采样时刻t′ n的浓度计算值为：
Figure PCTCN2022073490-appb-000007
其中，C r，0表示第r个胞外代谢物初始浓度值。
采样点可以根据生产实际来选取，简单操作是在代谢通量V t的每个计算时间间隔内进行一次采样观测，即Z＝N-1，也可以在计算时间间隔内进行多次采样；一般情形下代谢通量V t的第一个和最后一个计算时间点应分别对应于实际采样观测的初始和最终时刻。
各采样时刻的代谢通量估计：
在采样观测时间点t′ n，各胞外代谢物的实际采样值为C′ r，n，则浓度计算值与采样值的偏差为：
Figure PCTCN2022073490-appb-000008
其中,
Figure PCTCN2022073490-appb-000009
为胞外代谢物r在采样时刻n的浓度误差项方差。
以
Figure PCTCN2022073490-appb-000010
表示各个计算时刻的d维自由通量，通过优化算法寻找满足式(5)的U，使得偏差Φ最小，即求解
Figure PCTCN2022073490-appb-000011
获得最优的自由通量U，也即各t D时刻的自由通量u D，再利用式(2)估计代谢通量V t。
简单的优化算法是直接由式(6)对U求导数，即
Figure PCTCN2022073490-appb-000012
得到优化变量U＝H -1·J，其中矩阵H和矩阵J由下式给定：
Figure PCTCN2022073490-appb-000013
Figure PCTCN2022073490-appb-000014
式中W r是由
Figure PCTCN2022073490-appb-000015
构成的对角矩阵。
工程实际中可增加初始浓度C R0、代谢通量计算时刻t D为优化变量，利用优化算法求解
Figure PCTCN2022073490-appb-000016
获得最优的自由通量U、初始浓度C 0以及时刻t D。
建立运行状况监测模型及指标：
扩展样本数据集
生产中已有的I个正常批次数据，其中第i批次生产变量(包括温度、pH通风率、搅拌功率、底物流加速率等)数据矩阵为X i∈R N×O，i＝1，...，I，N为采样观测点数，O为生产变量数。根据估计所得的代谢通量V t，选取其中全部或部分代谢通量，若选取L个胞内外交换代谢通量在第i批次的数据构成集合Y i∈R N×L，则该批次的生产变量数据与胞内外交换代谢通量合并构成扩展样本数据集G i＝[X i，Y i] T∈R N×H，其中H＝O+L。全部I个批次的发酵过程扩展样本数据集则为
Figure PCTCN2022073490-appb-000017
按批次方向展开得G∈R I×N·H。
数据集的标准化处理
将扩展样本数据集标准化变换为均值为0、方差为1的标准训练数据集，近似服从多维正态分布：
Figure PCTCN2022073490-appb-000018
其中，mean(G)为G中同一时刻各个批次同一变量的均值，即多批次正常操作下的平均运行轨迹，std(G)是G中同一时刻各个批次同一变量的标准 差。
监测模型及控制限的求取
构造标准数据集G train的协方差矩阵
Figure PCTCN2022073490-appb-000019
利用主元分析建立监测模型，先求取其特征值及其特征向量，按累计方差贡献率从大到小，选取A个主元特征值所对应的特征向量q 1，…，q A,构成载荷矩阵Q＝[q 1，…，q A]，进而计算G train的各批次得分向量和残差向量，全部I批次得分向量和残差向量分别构成得分矩阵和残差矩阵，据此求取常用的T 2控制限和SPE控制限。
监测模型也可以采用多阶段多向主元分析、多向核主元分析和多类支持向量机等方法。
在线实时运行状况监测：
当前时刻数据处理
采集生产过程当前时刻n的生产变量数据X test，n∈R n×O，通过估计获得的代谢通量Y test，n∈R n×L，将生产变量和代谢通量合并得到测试数据集G test，n∈R n×H；以训练数据集中相应时刻的平均值和标准差对G test，n∈R n×H进行标准化处理。
未来时刻数据估计
当前时刻之后的n+1至批次结束N时刻的数据，采用预估的方法进行填充，用标准化的G test，n中第n时刻数据值进行填充，得到G′ test，n∈R N×H，再按时间方向展开为G″ test，n∈R 1×N·H。也可以直接用前若干批次对应时刻的均值数据，或考虑当前及过去时刻实际数据与对应均值数据之差，预估填充未来数据。
计算当前时刻T 2统计量和SPE统计量
根据标准训练数据集G train的载荷矩阵Q计算当前n时刻得分向量t test，n和残差向量e test，n:
Figure PCTCN2022073490-appb-000020
按下式计算n时刻T 2统计量和SPE统计量:
Figure PCTCN2022073490-appb-000021
Figure PCTCN2022073490-appb-000022
其中，Ψ∈R A×A是由G train的第i批次得分向量中各个元素构成的对角矩阵。
过程运行状况监测
若T 2和SPE n低于统计量控制限，则过程运行正常；否则预警过程有异常或故障。
为了验证本实施例的利用细胞代谢网络监测生物制造过程的方法的效果，将方法应用于青霉素生产过程运行监测：
青霉素发酵代谢网络共包含66个代谢通量(其中胞内外交换通量5个)和49种胞内代谢物，葡萄糖、青霉素、菌体生物质、氧气及二氧化碳等5种胞外代谢物可检测，考虑反应温度、pH值、通风率、搅拌功率、底物流加速率和冷却水流加速率等生产工艺变量和胞内外交换代谢通量共11个，选取30批次、每批次800采样点，形成扩展样本数据集。
按本发明方法完成代谢通量的胞外代谢物浓度计算和各采样时刻的代谢通量估计，图3给出胞内外交换代谢通量的估计值，由此扩展样本数据集不仅包含了生物发酵过程可检测的生产工艺变量信息，也包含了细胞内部不可测的代谢信息，更全面地反映生产过程特征。
针对青霉素发酵过程由通风、搅拌及底物流加引起的典型故障，图4给出了本发明在线监测结果：
其中(a)表示的过程监测只采用了生产工艺变量，而(b)则同时结合代谢 网络信息。从T 2监测图中可以发现，当故障发生时，本发明所提出的方法能够及时准确地监测故障，而基于生产变量监测有一定时间滞后。
本领域内的技术人员应明白，本申请的实施例可提供为方法、系统、或计算机程序产品。因此，本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请是参照根据本申请实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
显然，上述实施例仅仅是为清楚地说明所作的举例，并非对实施方式的限定。对于所属领域的普通技术人员来说，在上述说明的基础上还可以做出其它不同形式变化或变动。这里无需也无法对所有的实施方式予以穷举。而 由此所引伸出的显而易见的变化或变动仍处于本发明创造的保护范围之中。

Claims (10)

  1. 一种利用细胞代谢网络监测生物制造过程的方法，其特征在于：包括如下步骤：基于代谢网络进行通量均衡分析，得到反映细胞生长与繁殖过程的代谢通量；将细胞内代谢通量与生产过程变量一起构成训练数据集，建立生物发酵过程监测模型及统计量控制限；基于监测模型，利用实时采集的生产过程变量和代谢通量计算值进行运行状况在线监测。
  2. 根据权利要求1所述的利用细胞代谢网络监测生物制造过程的方法，其特征在于：所述通量均衡分析过程包括：建立通量均衡方程；建立胞外代谢物浓度与代谢通量的动态关系；根据多批次胞外代谢物浓度采样值估计代谢通量。
  3. 根据权利要求2所述的利用细胞代谢网络监测生物制造过程的方法，其特征在于：所述代谢通量的随时间变化，采集相邻两个时间点之间的代谢通量变化规律对代谢通量进行瞬态特征描述，包括常数、线性函数、二次函数。
  4. 根据权利要求2所述的利用细胞代谢网络监测生物制造过程的方法，其特征在于：多批次胞外代谢物浓度采样值根据生产实际来选取，在计算时间间隔内采样一次或多次采样，再进行曲线拟合。
  5. 根据权利要求2所述的利用细胞代谢网络监测生物制造过程的方法，其特征在于：在估计代谢通量时，通过求导数、线性规划、二次规划的优化算法求得最优代谢通量。
  6. 根据权利要求1所述的利用细胞代谢网络监测生物制造过程的方法，其特征在于：建立生物发酵过程监测模型包括：将生产过程变量和细胞内代谢通量合并构成扩展样品数据集；对样品数据集进行标准化处理生成标准样品数据集；利用主元分析建立监测模型；确定监测统计指标的控制限。
  7. 根据权利要求6所述的利用细胞代谢网络监测生物制造过程的方法， 其特征在于：在扩展样品数据集时，所述细胞内代谢通量选取全部代谢通量或胞内外交换代谢通量或内部代谢通量。
  8. 根据权利要求6所述的利用细胞代谢网络监测生物制造过程的方法，其特征在于：所述主元分析包括一般主元分析、多阶段主元分析、核主元分析以及支持向量机方法。
  9. 根据权利要求1所述的利用细胞代谢网络监测生物制造过程的方法，其特征在于：所述运行状况在线监测包括：采集当前批次当前时刻的胞外代谢物浓度和生产过程变量；由胞外代谢物浓度估计获得的代谢通量；将生产过程变量与代谢通量合并构成测试数据集；对测试数据集进行标准化处理生成测试标准数据集；预估填充未来时刻的测试数据；计算统计标准，并与统计量控制限比较。
  10. 根据权利要求9所述的利用细胞代谢网络监测生物制造过程的方法，其特征在于：预估填充未来时刻的测试数据的方法包括直接用前若干批次对应时刻的均值数据、或考虑本批次当前及过去时刻实际数据与对应均值数据之差，预估填充未来数据。
PCT/CN2022/073490 2021-08-18 2022-01-24 利用细胞代谢网络监测生物制造过程的方法 Ceased WO2023019883A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US18/396,157 US20240273262A1 (en) 2021-08-18 2023-12-26 Method for monitoring biomanufacturing process using metabolic network of cells

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110950873.9A CN113838530B (zh) 2021-08-18 2021-08-18 利用细胞代谢网络监测生物制造过程的方法
CN202110950873.9 2021-08-18

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US18/396,157 Continuation US20240273262A1 (en) 2021-08-18 2023-12-26 Method for monitoring biomanufacturing process using metabolic network of cells

Publications (1)

Publication Number Publication Date
WO2023019883A1 true WO2023019883A1 (zh) 2023-02-23

Family

ID=78960867

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/073490 Ceased WO2023019883A1 (zh) 2021-08-18 2022-01-24 利用细胞代谢网络监测生物制造过程的方法

Country Status (3)

Country Link
US (1) US20240273262A1 (zh)
CN (1) CN113838530B (zh)
WO (1) WO2023019883A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN120496067A (zh) * 2025-07-15 2025-08-15 青岛奥克生物开发有限公司 基于多参数的间充质干细胞培养监测系统
CN120866584A (zh) * 2025-07-29 2025-10-31 辽宁省农业科学院 一种秸秆低温高效促腐菌剂制备控制方法及系统

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113838530B (zh) * 2021-08-18 2022-07-15 江南大学 利用细胞代谢网络监测生物制造过程的方法
CN116130020B (zh) * 2023-02-27 2026-02-10 哈尔滨工业大学 基于代谢动力学-gru的有机固废堆肥过程nh3产生预测方法
CN116798521B (zh) * 2023-07-19 2024-02-23 广东美赛尔细胞生物科技有限公司 免疫细胞培养控制系统的异常监测方法及系统
CN119724368B (zh) * 2025-02-26 2025-07-04 深圳诺普信作物科学股份有限公司 一种基于菌株繁殖特性分析的微生物菌剂制备控制方法
CN120257129B (zh) * 2025-06-04 2025-08-12 北京盛拓达生物技术有限公司 一种发酵过程的异常状态预警方法、系统、设备及介质
CN120600095B (zh) * 2025-06-10 2025-12-23 北京德量源生物科技有限公司 用于单细胞蛋白生产的双碳源混合发酵代谢调控系统
CN120766283A (zh) * 2025-07-10 2025-10-10 陕西开盈锐门生物科技有限公司 基于物联网的干细胞提取智能监测方法及系统

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190221282A1 (en) * 2018-01-18 2019-07-18 Tata Consultancy Services Limited Method and system for analyzing metabolic state of a cell by measuring concentrations of metabolites
CN111210867A (zh) * 2020-02-10 2020-05-29 江南大学 基于代谢分析及酶调控的微生物生长状态估计方法
CN111341382A (zh) * 2020-02-20 2020-06-26 江南大学 赖氨酸生物制造中宏观动力学与细胞代谢通量耦合建模方法
CN112119306A (zh) * 2017-12-29 2020-12-22 豪夫迈·罗氏有限公司 细胞培养物的代谢状态的预测
CN113838530A (zh) * 2021-08-18 2021-12-24 江南大学 利用细胞代谢网络监测生物制造过程的方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112119306A (zh) * 2017-12-29 2020-12-22 豪夫迈·罗氏有限公司 细胞培养物的代谢状态的预测
US20190221282A1 (en) * 2018-01-18 2019-07-18 Tata Consultancy Services Limited Method and system for analyzing metabolic state of a cell by measuring concentrations of metabolites
CN111210867A (zh) * 2020-02-10 2020-05-29 江南大学 基于代谢分析及酶调控的微生物生长状态估计方法
CN111341382A (zh) * 2020-02-20 2020-06-26 江南大学 赖氨酸生物制造中宏观动力学与细胞代谢通量耦合建模方法
CN113838530A (zh) * 2021-08-18 2021-12-24 江南大学 利用细胞代谢网络监测生物制造过程的方法

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN120496067A (zh) * 2025-07-15 2025-08-15 青岛奥克生物开发有限公司 基于多参数的间充质干细胞培养监测系统
CN120866584A (zh) * 2025-07-29 2025-10-31 辽宁省农业科学院 一种秸秆低温高效促腐菌剂制备控制方法及系统

Also Published As

Publication number Publication date
CN113838530B (zh) 2022-07-15
US20240273262A1 (en) 2024-08-15
CN113838530A (zh) 2021-12-24

Similar Documents

Publication Publication Date Title
CN113838530B (zh) 利用细胞代谢网络监测生物制造过程的方法
US11795516B2 (en) Computer-implemented method, computer program product and hybrid system for cell metabolism state observer
CN112119306B (zh) 细胞培养物的代谢状态的预测
US20250068794A1 (en) Method for Predicting Outcome of an Modelling of a Process in a Bioreactor
Charaniya et al. Mining manufacturing data for discovery of high productivity process characteristics
CN101308385B (zh) 基于二维动态核主元分析的非线性过程故障检测方法
Jiang et al. Weighted kernel principal component analysis based on probability density estimation and moving window and its application in nonlinear chemical process monitoring
CN117229905B (zh) 一种生物饲料发酵管控方法及系统
CN115175982B (zh) 用于模拟细胞培养过程的计算机执行的方法、计算机程序产品和系统
CN101446831B (zh) 一种分散的过程监测方法
CN101964021B (zh) 基于递归核主元分析的青霉素发酵过程故障监测方法
Yang et al. A hybrid model-based fault detection strategy for air handling unit sensors
CN114169254A (zh) 基于短期建筑能耗预测模型的异常能耗诊断方法及系统
Zhang et al. Dynamical process monitoring using dynamical hierarchical kernel partial least squares
CN115407753B (zh) 一种多变量加权集成学习的工业故障诊断方法
CN109144039A (zh) 一种基于时序扩展和邻域保持极限学习机的间歇过程故障检测方法
CN113253682A (zh) 非线性化工过程故障检测方法
CN110046377B (zh) 一种基于异构相似度的选择性集成即时学习软测量建模方法
CN103559401B (zh) 基于半监督主元分析的故障监控方法
González-Martínez et al. Metabolic flux understanding of Pichia pastoris grown on heterogenous culture media
CN117932377A (zh) 基于聚类学习的冷轧铝板带材厚度控制方法及相关设备
CN116776252A (zh) 一种改进Mallow's Cp变量选择的工业过程软测量方法和系统
Wang et al. Weak fault monitoring method for batch process based on multi-model SDKPCA
CN116469478A (zh) 一种基于深度学习的青霉素发酵过程在线软测量方法和装置
CN120949587A (zh) 一种提高菌体活性与稳定性的多阶段发酵方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22857205

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 22857205

Country of ref document: EP

Kind code of ref document: A1