Search

CN-122024813-A - Virus vector production control method based on virtual cells

CN122024813ACN 122024813 ACN122024813 ACN 122024813ACN-122024813-A

Abstract

The application relates to the technical field of intelligent control of medicine production, and discloses a virus vector production control method based on virtual cells, which comprises the steps of constructing a virtual cell (HEK 293) model of an integrated multi-sub model based on bioreactor real-time data, laboratory material quality data and equipment static data; the method comprises the steps of coupling transfection or infection dynamics, cell population and bioreactor transfer model, establishing an upstream process digital twin body, guiding a process optimization target, obtaining upstream culture and transfection parameters through an optimization algorithm, and finally generating control instructions or parameter settings of actual viral vector production based on the optimization parameters. The method realizes the digital and intelligent regulation and control of the production process, improves the quality stability and production efficiency of the product, and reduces the cost and the amplification risk.

Inventors

  • ZHOU LU
  • GAO XINYU

Assignees

  • 神拓生物技术(杭州)有限公司

Dates

Publication Date
20260512
Application Date
20260212

Claims (12)

  1. 1. A virus vector production control method based on virtual cells is characterized by comprising the following steps: s1, constructing a virtual cell model corresponding to virus vector production cells based on historical physiological parameter data, wherein the virtual cell model is integrated with a genome scale metabolic network model, a virus assembly dynamics sub-model and a cell stress response sub-model, and the historical physiological parameter data is derived from real-time process time sequence data of a bioreactor, material and quality attribute data from a laboratory information management system and static configuration data of production equipment; s2, based on the virtual cell model, coupling transfection or infection dynamics model, cell population model and bioreactor transfer model, establishing an upstream process digital twin body; S3, based on the upstream process digital twin body, performing an optimization algorithm by taking a preset process optimization target as a guide to obtain optimized process parameters, wherein the optimized process parameters comprise upstream culture parameters and transfection parameters; s4, generating instructions or parameter settings for controlling the actual viral vector production process based on the optimized process parameters.
  2. 2. The method for controlling production of a virtual cell-based viral vector according to claim 1, The virtual cell model is constructed based on the historical physiological parameter data, specifically, a preset genome scale metabolic network model is loaded; calibrating the metabolic exchange reaction flux constraint of the preset genome scale metabolic network model based on the historical cell physiological parameter data to obtain a calibrated metabolic network model; And logically connecting the calibrated metabolic network model with a preset virus assembly dynamics sub-model and a cell stress response sub-model to form the virtual cell model.
  3. 3. The method for controlling production of a virtual cell-based viral vector according to claim 2, The creation of an upstream process digital twin, specifically, instantiating the virtual cell model into a plurality of virtual cell models to construct the cell population model; Coupling the cell population model with the transfection or infection dynamics model and the bioreactor transfer model, wherein the environmental parameters calculated by the bioreactor transfer model act on the cell population model, and the metabolism and production state summarized by the cell population model update the material field of the bioreactor transfer model; In the actual production process of the virus vector, based on real-time process time sequence data, adopting a data assimilation algorithm to dynamically adjust internal state parameters of the upstream process digital twin body.
  4. 4. A method for controlling production of a virtual cell-based viral vector according to claim 3, The data assimilation algorithm is specifically an extended Kalman filtering algorithm, the state vector of the upstream process digital twin body comprises cell density, nutrient concentration and metabolic byproduct concentration, and the real-time process time sequence data is used as an observation vector for updating the estimated value of the state vector.
  5. 5. The method for controlling production of a virtual cell-based viral vector according to claim 1, The execution optimization algorithm is specifically a Bayesian optimization algorithm, and specifically comprises the following steps: training a Gaussian process regression model as a proxy model of an optimization target based on the historical experimental data set; selecting a next process parameter combination to be evaluated based on the proxy model and the acquisition function within a parameter search space defined by process constraints; inputting the technological parameter combination to be evaluated into the upstream technological digital twin body to carry out virtual production simulation, so as to obtain a predicted technological result; adding the technological parameter combination to be evaluated and the corresponding prediction technological result into the historical experimental data set, and updating the agent model; And iteratively executing the steps of selecting, simulating and updating until the stopping condition is met, and selecting the optimized process parameters from all the evaluated process parameter combinations.
  6. 6. The method for controlling production of a virtual cell-based viral vector according to claim 1, After the step S3 and before the step S4, the method further includes: Optimizing the downstream purification process parameters based on the downstream unit operation digital model and a multi-objective optimization algorithm to obtain downstream optimization process parameters; wherein the downstream unit operation digital model comprises a chromatography model based on a multicomponent adsorption dynamics equation and a membrane filtration model based on a blocking mechanism.
  7. 7. The method for controlling production of a virtual cell-based viral vector according to claim 6, The multi-objective optimization algorithm is a non-dominant ranking genetic algorithm, and its optimization objectives include at least two of virus recovery, impurity residual levels, and production costs.
  8. 8. The method of claim 1, wherein the step of determining the position of the substrate comprises, After the step S3 and before the step S4, the method further includes: taking the optimized process parameters as input, and performing virtual amplified production simulation in the scaled upstream process digital twin body based on an amplification criterion and target production scale equipment parameters; Comparing the virtual amplified production simulation result with the small-scale simulation result, and identifying the predicted variation of the key performance index; Based on a preset risk threshold, judging whether the predicted variation forms a process amplification risk or not, and generating a corresponding risk assessment result.
  9. 9. The method for controlling production of a virtual cell-based viral vector according to claim 1, further comprising: Constructing a quality attribute prediction model, wherein the quality attribute prediction model takes production process data and/or process parameters as input and takes a key quality attribute prediction value of a virus vector as output; In the actual production process control, inputting real-time or staged production process data into the quality attribute prediction model to obtain a real-time key quality attribute prediction value; And comparing the real-time key quality attribute predicted value with a preset product release standard to generate a real-time release test conclusion.
  10. 10. The method for controlling production of a virtual cell-based viral vector according to claim 1, The virus vector production cell is in particular HEK293 cell, or other mammalian cell or derived stable production cell line for virus vector production.
  11. 11. The method for controlling production of a virtual cell-based viral vector according to claim 10, The other mammalian cells include at least any one of the following: HEK293T cells, CHO cells, per.c6 cells, vero cells, BHK cells, CAP cells or MDCK cells.
  12. 12. A virtual cell-based viral vector production control system, comprising: the system comprises a virtual cell model construction module, a virtual cell model analysis module and a control module, wherein the virtual cell model construction module is used for constructing a virtual cell model based on historical physiological parameter data, and the virtual cell model is integrated with a genome scale metabolic network model, a virus assembly dynamics sub-model and a cell stress response sub-model, wherein the historical physiological parameter data is derived from real-time process time sequence data of a bioreactor, material and quality attribute data from a laboratory information management system and static configuration data of production equipment; The upstream twin body construction module is used for constructing an upstream process digital twin body based on the virtual cell model, the coupled transfection or infection dynamics model, the cell population model and the bioreactor transfer model; The process optimization module is used for executing an optimization algorithm based on the upstream process digital twin body by taking a preset process optimization target as a guide to obtain optimized process parameters, wherein the optimized process parameters comprise upstream culture parameters and transfection parameters; And the production control instruction generation module is used for generating instructions or parameter settings for controlling the actual viral vector production process based on the optimized process parameters.

Description

Virus vector production control method based on virtual cells Technical Field The application relates to the technical field of intelligent control of drug production, in particular to a virus vector production control method based on virtual cells. Background With the rapid development of gene therapy technology, lentiviruses and adeno-associated viruses have become core tools for gene delivery, and the market demand thereof continues to proliferate with the progress of clinical transformation and commercialization. However, the large-scale production of viral vectors still faces a number of technical bottlenecks to be solved, which severely restricts the accessibility of gene therapy drugs. The development of the traditional viral vector production process is highly dependent on the step-by-step trial and error of laboratory scale, and when the process is enlarged from the laboratory scale to pilot scale and commercial production scale, the problems of process amplification effect failure such as titer reduction, impurity increase, recovery rate reduction and the like often occur. HEK293 cells as production cores are susceptible to multiple factors such as substrate concentration, plasmid quality, culture conditions and the like in transfection efficiency, metabolic state and virus assembly rate, and the fluctuation among production batches is remarkable. Meanwhile, the quality detection means is more lagged behind the production flow, real-time early warning and accurate regulation and control in the process are difficult to realize, and production risks cannot be avoided timely. In addition, the cost of key raw materials such as plasmid DNA, special culture medium, chromatography purification consumable materials and the like is high, and the whole production cost of the viral vector is further increased. The problems of difficult process amplification, poor batch stability, lag in quality control, high production cost and the like become key barriers for impeding the large-scale application of gene therapy technology, and development of an intelligent and digital production control scheme for realizing the accurate prediction and efficient regulation of the whole production flow of virus vectors is needed. Disclosure of Invention The present application is directed to a method for controlling production of a viral vector based on virtual cells, which solves the problems set forth in the background art. According to a first aspect of the present application, there is provided a method for controlling production of a viral vector based on virtual cells, comprising the steps of: s1, constructing a virtual cell model based on historical physiological parameter data, wherein the virtual cell model is integrated with a genome scale metabolic network model, a virus assembly dynamics sub-model and a cell stress response sub-model, and the historical physiological parameter data is derived from real-time process time sequence data of a bioreactor, material and quality attribute data from a laboratory information management system and static configuration data of production equipment; s2, based on the virtual cell model, coupling transfection or infection dynamics model, cell population model and bioreactor transfer model, establishing an upstream process digital twin body; S3, based on the upstream process digital twin body, performing an optimization algorithm by taking a preset process optimization target as a guide to obtain optimized process parameters, wherein the optimized process parameters comprise upstream culture parameters and transfection parameters; s4, generating instructions or parameter settings for controlling the actual viral vector production process based on the optimized process parameters. Preferably, the virtual cell model is constructed based on the historical physiological parameter data, specifically, a preset genome scale metabolic network model is loaded; calibrating the metabolic exchange reaction flux constraint of the preset genome scale metabolic network model based on the historical cell physiological parameter data to obtain a calibrated metabolic network model; And logically connecting the calibrated metabolic network model with a preset virus assembly dynamics sub-model and a cell stress response sub-model to form the virtual cell model. Preferably, the establishing an upstream process digital twin body, specifically, instantiating the virtual cell model into a plurality to construct the cell population model; Coupling the cell population model with the transfection or infection dynamics model and the bioreactor transfer model, wherein the environmental parameters calculated by the bioreactor transfer model act on the cell population model, and the metabolism and production state summarized by the cell population model update the material field of the bioreactor transfer model; In the actual production process of the virus vector, based on real-time process time sequence data