Model-based analysis tools for bioreactors

Abstract

A method and system for analyzing biomass in a bioreactor (3) via a computer (2) with system software (5), the bioreactor (3) having at least one sensor (6) for measuring biomass and having a data connection to the computer (2) managed by a data interface provided by the system software (5), the system software (5) providing a data conversion model (8) for analyzing real-time raw data on dielectric constant measured by the at least one sensor (6) and transmitted to the computer (2) to calculate specific cellular parameters of the cells in the biomass.

Description

[Technical field]

[0001] The invention described herein discloses a method for operating in situ analytical tools within a bioreactor using computer-assisted, physics-based models.

[0002] Technical Field The present invention relates to the technical field of continuous biopharmaceutical processing. [Background technology]

[0003] The pharmaceutical industry's quality approach is focused on improving and increasing productivity in the manufacture of biochemical compounds. This requires the use of complex bioprocesses with real-time monitoring integrated within the production line. Inline analysis can enable the automation of processes, which can then be optimized by saving significant time and materials. Currently, there is a wide range of sensors and offline technologies on the market that can monitor key variables in cell culture, such as biomass, radius, nutrient content and metabolic indicators, as well as key parameters in bioprocesses, but most of them have not been translated into in situ sensors.

[0004] Therefore, converting analytical tools into in situ sensors is a current exploratory trend aimed at improving the quality of these measurements. Furthermore, thanks to these optimized sensors, called process analytical tools (PATs), the conditions of continuous or discontinuous cell cultures can be adjusted in real time, thanks to physical measurements converted into quantitative and qualitative information through models. This adaptation to inline sensors has many advantages, such as no cleaning steps, reduced system downtimes, no need for clean rooms, and reduced costs.

[0005] Another property is the conversion from multi-use (MU) to single-use (SU) sensors, which offers similar advantages, especially the elimination of the cleaning step. Unfortunately, SU sensors have major problems with calibration, which cannot be performed before the system is installed.

[0006] Therefore, these process sensors and analytical tools require specific and complex calibration models based on large amounts of data to address these challenges.

[0007] In summary, there are four main issues with the known state of the art described above. 1. Usually, when first used in a specific application and process, PAT only provides raw data and does not directly give parameter information and measurements such as viable cell density, glucose concentration, etc. For example, dielectric spectroscopy gives quantitative medium permittivity data but not viable cell density data. Therefore, complex conversion or calibration models need to be developed.

[0008] 2. The data-driven calibration model for PAT is the preferred model since there are no other approaches currently implemented and used in this application field. Data-driven calibration models for PAT require several cell culture runs and large amounts of data to give parameter measurements with acceptable accuracy and measurement tolerances.

[0009] 3. Among the difficulties in converting a multi-use sensor or PAT to a single-use version, the main difficulty with SU variations is the specific calibration of these, which is very different from the calibration of MU sensors. On the one hand, process MU sensors can be calibrated offline just before a run, whereas process SU sensors require pre-calibration data given by the supplier. Data-driven calibration models for analytical tools cannot be transferred from one MU to another MU or SU PAT, because parts of the model depend on the probe, such as specific sensitivity, internal factory coefficients, etc.

[0010] 4. Scaling up from, say, a 3L bioreactor to a much larger one of, say, 2kL, poses a challenge for in situ analysis, as these models are data-driven based. These data can be sensitive to culture conditions that vary greatly depending on the size of the bioreactor and the amount of mixing, sparging, etc. Summary of the Invention [Problem to be solved by the invention]

[0011] The problem of this patent application is therefore to find a way to use analytical tools in bioreactors that can overcome the known limitations of the prior art.

[0012] This problem is solved by a method for analyzing biomass in a bioreactor via a computer with system software, the bioreactor having at least one sensor for measuring biomass and having a data connection to the computer managed by a data interface provided by the system software, and the system software analyzes real-time raw data on dielectric constant measured by the at least one sensor and transmitted to the computer to provide a data conversion model for calculating specific cellular parameters of cells in the biomass.

[0013] The aim of the present invention is to transform a sensor, in this case a capacitance probe, and integrate dielectric spectroscopy into a true biomass probe to provide qualitative and quantitative information on cellular parameters such as radius and viable cell density. More importantly, the probe operates in real time to provide raw data, reduces calibration efforts, and accommodates the variability of multi-use or single-use probes. This approach solves the four problems described one by one.

[0014] Issues 1 and 2: The physics-based model is available from the first use of the probe and does not require machine learning and / or model building for the model parameters and coefficients because this data is obtained from the probe measurements, estimated from offline measurements, or leveraged from the literature. The physics-based model is based on equations that describe cells as "dielectrics" and therefore does not require prior calibration based on large amounts of data or old cell culture runs. It can use real-time physical values ​​obtained from the probe.

[0015] Question 3: The physics-based model is sensor independent and does not require factory calibration, so the model can self-calibrate with used sensors. The parameters extracted from the equations are obtained from cells considered as dielectrics, so the model can be transferred from one multi-use probe to another MU probe or to a single-use probe.

[0016] Question 4: Although the physics-based model is cell line independent, the cells have a shape that is modeled in the model, and in fact, since the cells are considered as dielectrics, their biochemical specificity is not a root cause of interference in the model.

[0017] Cell membrane capacitance C m and internal conductivity σ i is calculated from the offline analysis and allows the model to be periodically adjusted while providing qualitative information of the cells.

[0018] Preferred further developments of this process include, for example, but are not limited to, the following: In addition to simply using physics-based data models to analyze real-time raw data, a data-driven machine learning approach is used for data transformation models to achieve hybrid data transformation models with improved accuracy. At least one sensor measures the amplitude of the dielectric constant at various excitation frequencies as real-time raw data.

[0019] The computer calculates the cell dimensions in the form of its radius or diameter and the viable cell density (VCD) as cell parameters taking into account predefined parameter values ​​of cell membrane capacitance and internal conductivity. Data are discontinuously adjusted based on sampling and off-line analysis of cell membrane capacitance and internal conductivity. After the end of each measurement turn, the average values ​​of the cell membrane capacitance and internal conductivity are calculated via offline analysis and used for the next measurement turn instead of the previously defined parameter values.

[0020] Another solution to this problem is an automated system for analyzing biomass, comprising a bioreactor with at least one sensor for measuring biomass, a computer connected to the at least one sensor, and system software running on the computer with a data interface managing the connection to the at least one sensor and providing a data conversion model, the system software being arranged to carry out the aforementioned method.

[0021] Preferred further developments of the automated system include, for example, but are not limited to, the following: At least one sensor is a capacitance probe integrated with dielectric spectroscopy. The software includes specific software modules implemented between the smart dielectric spectroscopy probe and the data interface, allowing real-time raw data processing with embedded models. At least one sensor is a disposable, single-use sensor. · The computer is the single control unit that runs the system software and the data transformation model.

[0022] The computers include a first computer connected to the at least one sensor and running system software that controls the bioreactor and has a data interface that manages the connection to the at least one sensor, and a second computer at a remote location that provides a data conversion model and uses a connection to the first computer via the data interface. The data conversion model can be used for separate sensors, independent of whether at least one of the sensors is a single-use or multi-use probe, i.e. the model can be used for one or more sensors that are multi- or single-use. [Means for solving the problem]

[0023] The method according to the invention and the automation system 1 including the software 5, as well as advantageous functional developments thereof, are explained in more detail below using at least one preferred exemplary embodiment and with reference to the associated drawings, in which elements corresponding to one another are provided with the same reference numbers. [Brief description of the drawings]

[0024] [Figure 1] FIG. 1 shows a schematic diagram of the automated bioreactor system used. [Figure 2a] FIG. 2 shows an easy to understand schematic diagram of various preferred embodiments of the model used. [Figure 2b] FIG. 2b shows an easy to understand schematic diagram of various preferred embodiments of the model used. [Diagram 3] FIG. 3 shows the resulting curves of viable cell density (VCD). [Figure 4] FIG. 4 shows the resulting curve of radius (R). [Diagram 5] FIG. 5 shows the average values ​​of cell membrane capacitance and internal conductivity. [Figure 6] FIG. 6 shows the respective resulting curves of viable cell density (VCD) comparing single-use and multi-use probes. [Figure 7]FIG. 7 shows the resulting curves in radius (R) representation comparing single-use and multi-use probes. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0025] Figure 1 shows an example of an automated bioreactor system 1 for use in the present invention. The system comprises the bioreactor 3 itself, which contains a biomass including a cell culture, its control unit 2, a biomass sensor 6 connected to the bioreactor 3, and system software 5 executed by the control unit 2, which calculates, using specific data models 8, certain cellular parameters of the cells in the biomass by analyzing real-time raw data on the dielectric constant measured by at least one sensor 6 and transmitted to the control unit 2.

[0026] The control unit 2 is preferably a standard computer suitable for controlling the bioreactor 3. Another option is a microcontroller or processor integrated in a device built into the bioreactor 3. It could also be a standard or industrial personal computer or server, or any other suitable device, especially if the local control unit 2 provides the data model 8 itself, which would require a higher processing power than is typically provided by a microcontroller. In another preferred embodiment, the data model 8 is provided by a suitable separate computer in a remote location via a data network using a cloud-based service.

[0027] The data model8 is preferably a phenomenological Cole-Cole model8 that converts the raw real-time dielectric constant data into indices of viable cell density (VCD) and mean cell culture radius (R). The Cole-Cole equation is based on the Debye equation (Debye, 1929) and reproduces the shape of the β dispersion by expressing the dielectric constant (ε) as a function of frequency (f), which can be written as:

number

[0028] where Δε is the amplitude of the distribution, f c is the characteristic frequency (the frequency at which ε is equal to half the value of Δε), α is the slope of the distribution, ε0 is the permittivity of free space, ε ∞ is the dielectric constant at high frequencies (usually above 1 MHz) [Opel et al., 2010].

[0029] Dielectric parameters Δε, f c , and α are calculated by INCYTE internal software (ArcAir, Hamilton) from the raw permittivity data each time a scan is performed. The Cole-Cole parameters can be linked to quantitative information about cells, such as the average cultured cell radius R, using the following equation:

number

[0030] Here, C m (F / m 2 ) and σ i (measured in S / m) are the average membrane capacitance and internal conductivity, respectively, of cells in culture. σ a The quantity (measured in S / m) represents the static medium conductivity and can be determined from the following equation:

number

[0031] where σ (measured in S / m) is the static suspension conductivity and p p is the predicted biomass volume fraction expressed as:

number

[0032] Finally, the viable cell density VCD is calculated from the assumption that cells in culture are spherical, and the single cell volume V can be expressed as:

number

number

[0033] The software 5 for providing and applying the Cole-Cole model 8 also includes a raw data conversion module. Its graphical user interface (GUI) 4 allows the user 7 to select the type of modeling to use for the calculations. The software 5 is preferably MATLAB® software (The MathWorks Inc), although other suitable software can be used. In this example, MATLAB® version 9.9.0.1570001 of 2020 was used.

[0034] Using Model 8 algorithm, r and VCD values ​​were calculated every minute. Two samples were taken each day to obtain average offline values ​​of cell radius and VCD. These were interpolated with a smoothing spline. The values ​​calculated by Model 8 were compared to the spline and the standard error prediction (SEP) was calculated as follows:

number

[0035] Computer software 5 is preferably integrated into the platform for monitoring the radius and VCD throughout the culture. Using this GUI 4, the user can m and σ i The user is prompted to input a file containing the theoretical values ​​for, as well as the raw permittivity values. Depending on the selected model 8, it is also possible to add a file containing values ​​determined offline with a Nova analyzer. The raw permittivity data can also be provided in real time by the biomass sensor 6 as an alternative option.

[0036] The calculated radius and VCD values ​​are compared to offline measurements performed on an automated cell culture analyzer, thereby testing the validity of the Cole-Cole model8 applied to cells in culture.

[0037] Certain software modules are preferably implemented in the system software between the smart dielectric spectroscopy probe and the software interface, allowing real-time raw data processing by the embedded model 8.

[0038] The following method steps provide a preferred example for using Model 8 with the highest accuracy. 1) The described purely physics-based Cole-Cole model 8 is used together with the probe 6 to provide real-time permittivity measurements at various excitation frequencies. Real-time means at the fastest one measurement every 6 seconds. The probe 6 is directly used as a biomass sensor 6 from the first field use, using cell-specific parameters taken from the literature, preferably cell membrane capacitance and internal conductivity. This can be used up to the first 2 or 3 days of cell culture in the bioreactor 3. Figures 3 and 4 show the resulting curves for indication of viable cell density (VCD) and radius (R).

[0039] 2) Discontinuous adjustment of the transformation model 8 based on sampling and offline analysis of cell membrane capacitance and internal conductivity. Model 8 starts its calculations at each sampling of these cell-specific parameters based on the following equations:

number

number

[0040] Figure 5 shows the mean values ​​of each of these two cell-specific parameters, which can be calculated after the run is completed and subsequently used in place of the parameter values ​​from the literature.

[0041] 3) As experimental data shows, the model is transferable to disposable, single-use sensors without any specific sensor tuning. Figures 6 and 7 show the resulting curves in terms of viable cell density (VCD) and radius (R), respectively.

[0042] In conclusion, it can be seen that the calibrated model 8 can be used with either the MU or SU probe 6 without the additional calibration step of the SU sensor that is normally required for typical process control sensors such as pH and dissolved oxygen, without losing the calibration-free feature of the present invention.

[0043] Since Model 8 is cell line independent and uses cells as dielectric, its scalability to characterize and monitor cell cultures from small to large bioreactors is evident. Improvement of the accuracy of Model 8 is achieved through hybrid models that combine data-driven approaches with physics-based Model 8. Figure 2 provides an easy to understand schematic diagram of the present invention encompassing various preferred embodiments of Model 8 used. [Explanation of symbols]

[0044] 1. Automated Bioreactor System 2 Control Unit / Computer 3. Bioreactor 4 User Interface 5. Software 6 Sensors / Probes 7 Users 8 Data Conversion Model (Cole-Cole)

Claims

1. A method for analyzing biomass in a bioreactor (3) via a computer (2) equipped with system software (5), the bioreactor (3) having at least one sensor (6) for measuring biomass and having a data connection to the computer (2) managed by a data interface provided by the system software (5), The system software (5) provides a data conversion model (8) for analyzing the real-time raw data on the dielectric constant measured by the at least one sensor (6) and transmitted to the computer (2) to calculate specific cellular parameters of the cells in the biomass. The method.

2. The method of claim 1, wherein a physics-based data model based on the Cole-Cole equations is used as the data transformation model (8).

3. 3. The method of claim 2, wherein in addition to analyzing the real-time raw data using a purely physics-based data model, a data-driven machine learning approach is used for the data transformation model (8) to achieve a hybrid data transformation model with improved accuracy.

4. 2. The method of claim 1, wherein at least one sensor (6) measures the amplitude of the permittivity at different excitation frequencies as real-time raw data.

5. 2. The method of claim 1, wherein the computer (2) calculates the cell dimensions in the form of its radius or diameter and the viable cell density (VCD) as cell parameters, taking into account predefined parameter values ​​of the cell membrane capacitance and the internal conductivity.

6. 6. The method of claim 5, wherein the data is discontinuously adjusted based on sampling and off-line analysis of cell membrane capacitance and internal conductivity.

7. 7. The method of claim 6, wherein after each measurement turn, average values ​​of the cell membrane capacitance and internal conductivity are calculated via offline analysis and used in the next measurement turn instead of the previously defined parameter values.

8. 8. An automated system for analyzing biomass, comprising: a bioreactor (3) equipped with at least one sensor (6) for measuring biomass; a computer (2) connected to the at least one sensor (6); and system software (5) running on the computer (2) equipped with a data interface managing the connection to the at least one sensor (6) and providing a data conversion model (8), the system being arranged to carry out the method according to any one of claims 1 to 7.

9. 9. The automated system according to claim 8, wherein at least one sensor (6) is a capacitance probe integrated with dielectric spectroscopy.

10. 10. The automated system according to claim 9, wherein the system software (5) comprises a specific software module implemented between the dielectric spectroscopy probe and the data interface, which allows real-time raw data processing in a data conversion model (8).

11. 9. The automated system of claim 8, wherein at least one sensor (6) is a disposable, single-use sensor.

12. 9. The automation system of claim 8, wherein the computer (2) is a single control unit that executes the system software (5) and the data conversion model (8).

13. 9. The automation system of claim 8, wherein the computer (2) comprises a first computer connected to at least one sensor (6), controlling the bioreactor (3) and running system software (5) with a data interface managing the connection to the at least one sensor (6), and a second computer at a remote location providing the data conversion model (8) and using a connection to the first computer via a data network to the data interface of the first computer.

14. 9. The automation system of claim 8, wherein the data conversion model (8) can be used for different sensors, independent of whether the at least one sensor (6) is a single-use or multi-use probe.

Images