Method and apparatus for chromatographic separation
Through a variety of chromatographic operation and computational grid analysis, the chromatographic analysis model calibration is simplified and the model parameters are quickly derived, which solves the complex and time-consuming problems of model calibration in the existing technology, and achieves the effect of rapid export under low load conditions.
Patent Information
- Application Number
- CN202380038003.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-05-05
- Filing Date
- 2023-04-27
- Publication Date
- 2024-12-13
AI Technical Summary
Model calibration process in existing chromatographic analysis is complex and time-consuming, especially when determining molecular specific parameters of isothermal equations. Traditional methods such as the Yamamoto method are effective under high load conditions, but may lead to misleading parameters under non-LMA type models. , and curve fitting requires a complete solution, resulting in a large time overhead.
By performing multiple chromatographic runs, recording the retention behavior data of the target molecules, generating calculation grids, selecting adsorption models, calculating division coefficient values and integrals, until the polymerization distance metric is minimized to determine adsorption model parameters, simplifying the model calibration process, and reducing experiments Number of times and computational complexity.
Rapidly export model parameters from low load density, reduce computational complexity and calibration time, suitable for non-LMA type models, and supports multi-mobile phase modifiers and nonlinear gradient elution experiments.
Smart Images
Figure CN119137477A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a method for performing chromatographic analysis, and more particularly to chromatographic separation of at least one target molecule using a specific chromatographic setting. Background Art
[0002] In order to perform chromatographic separations efficiently, chromatographic process development and / or process characterization and / or process optimization need to be performed. This involves tedious and time-consuming experimental work.
[0003] To improve this situation, mechanistic modeling of chromatography is increasingly used to address practical challenges of chromatography process development, process optimization, and process characterization through rapid and cost-effective computer simulation and optimization.
[0004] However, before a mechanistic model can be applied, it first needs to be calibrated for the chromatographic process that is about to occur. Chromatographic models of the prior art include concentration transport equations, which describe the movement of solutes through the interstitial volume of the chromatographic column. If the micropore volume filled with fluid is also involved, for example, when a packed bed of porous particles is used, the transport of solutes into and within the pores can also be simulated. The adsorption and desorption of solutes are described by the so-called isotherm equations, which are required for all types of chromatographic analysis except size exclusion chromatography (SEC). For example, ion exchange, hydrophobic interaction or affinity chromatography.
[0005] The biggest challenge in mechanistic modeling is determining the molecule-specific parameters for the isotherm equation, which describes the equilibrium state of the solute concentration both adsorbed and in solution.
[0006] Given the solute concentration in the mobile phase and the surrounding conditions (pH, ionic strength / charge, etc.), the equilibrium state is defined by how much solute can bind to the adsorbent. The equilibrium isotherm usually has a "linear range" at low loadings, where the concentration of molecules adsorbed is proportional to the concentration of molecules in the solute. At higher solute concentrations in the mobile phase, the curve flattens out as the adsorber capacity is approached. The isotherm equation describes the solute binding as a function of the mobile phase conditions (ionic strength, pH, etc.) and is molecule specific. Therefore, the isotherm equation includes multiple parameters whose values need to be determined by calibration experiments performed, for example, in the laboratory.
[0007] Some traditional approaches for model calibration use the so-called Yamamoto method. This method is preferably used for simple ion exchange chromatography (IEC) models, which separates the determination of the parameters affecting the linear region of the isotherm equation from the determination of the parameters affecting the nonlinear region. This reduces the dimensionality of the parameter determination problem and, as an additional advantage, the experiment requires only a small amount of sample in the linear region of the isotherm. Unfortunately, the parameter values determined by the Yamamoto method are only usable if the elution behavior under high load conditions can indeed be described by a simple stoichiometric displacement model (SDM) variant based on the law of mass action (LMA). If the binding behavior of the molecule does not follow the LMA / SDM mechanism, the determined parameters will be misleading and unusable. But even only slightly extended LMA-type SDM models, such as the IEC model from Mollerup, https: / / doi.org / 10.1016 / j.jbiotec.2007.05.036 , when solving the integral with the Yamamoto method, a closed-form formula may no longer be produced. Furthermore, new colloid models for chromatography are designed in an overly complex manner and are difficult to integrate. Therefore, the state of the art for non-LMA type models is to use curve fitting, which requires a complete solution of all model equations to calculate the chromatogram. Not only is this time consuming, but the parameter determination problem may become high-dimensional, resulting in a relatively large time overhead.
[0008] Therefore, there is a need for an improved method for performing chromatographic separations. In particular, there is a need for an improved method for performing chromatographic separations supported by mechanistic modeling. Purpose of the invention
[0009] An object of embodiments of the present invention is to provide a solution that alleviates or solves the above-mentioned disadvantages and problems. Summary of the invention
[0010] The above-mentioned objects and further objects are achieved by the subject matter described herein. Further advantageous implementation forms of the invention are also defined herein.
[0011] According to a first aspect of the present invention, the above-mentioned and other objects are achieved by a method, the method comprising: performing an experiment as a series of chromatographic runs, wherein the series of chromatographic runs comprises at least two chromatographic runs using different operating modes; recording retention behavior data of at least one target molecule for each of the chromatographic runs, wherein the retention behavior data at least indicates a measured elution point and a property of the measured elution point; generating at least one computational grid in at least one dimension, wherein the extent of the grid indicates a maximum range between a start point of the experiment and an end point of the experiment; selecting an adsorption model from a set of predetermined adsorption models, the adsorption model having initial candidate values for adsorption model parameters; using the selected adsorption model to calculate partition coefficient values for all points in the computational grid; calculating integrals of functions of the calculated partition coefficient values on at least two paths in the computational grid to obtain properties of the calculated elution points; calculating aggregate distance metrics from the properties of the calculated elution points to the properties of the corresponding measured elution points; selecting new candidate values for the adsorption model parameters; repeating the steps of calculating partition coefficient values, calculating integrals, calculating aggregate distance metrics, and selecting new candidate values until the aggregate distance metrics are minimized; using the selected values for the adsorption model parameters to determine parameters for the adsorption model that minimize the aggregate distance metrics.
[0012] In an embodiment according to the first aspect, the elution point measured comprises selection of a time point at elution and / or selection of a volume and / or selection of a mobile phase modifier concentration and / or a chromatographic run characteristic.
[0013] In an embodiment according to the first aspect, the breadth of the grid indicates the range of mobile phase modifier concentrations, wherein the function of the calculated partition coefficient value is an inverse function, wherein calculating the aggregate distance metric comprises: calculating the property of the calculated elution point by: using the inverse function of the calculated partition coefficient value to calculate the integral over the at least two paths in the calculation grid to obtain the property of the calculated elution point; calculating the aggregate distance metric between the property of the calculated elution point and the corresponding property of the measured elution point.
[0014] In an embodiment according to the first aspect, the method further comprises: selecting a chromatographic column model from a set of predetermined chromatographic column models; and determining candidate values for additional adsorption model parameters by fitting the retention behavior data to simulated chromatograms.
[0015] In an embodiment according to the first aspect, the extent of the computational grid indicates the duration in time of the series of chromatographic runs and the axial column dimension of the series of chromatographic runs, wherein calculating the aggregate distance metric comprises calculating the property of the calculated elution point by: selecting a chromatographic column model from a set of predetermined chromatographic column models; calculating a propagation velocity value of the at least one target molecule from an inlet of the column to an outlet of the column using the selected chromatographic column model and a partitioning coefficient value calculated on the computational grid; and simulating the retention behavior data of the at least one target molecule by: calculating the property of the calculated elution point by calculating an integral of the propagation velocity to obtain at least one trajectory through the computational grid, wherein the aggregate distance metric (D) is calculated between the property of the calculated elution point and the corresponding property of the measured elution point.
[0016] In an embodiment according to the first aspect, the method further comprises selecting candidate values for additional adsorption model parameters by fitting the retention behavior data to simulated chromatograms.
[0017] In an embodiment according to the first aspect, the method further comprises: calibrating a mechanistic model using the determined values of the adsorption model parameters, the chromatographic column model parameters and the selected candidate values of the additional adsorption model parameters; and simulating a chromatographic run using the calibrated mechanistic model to obtain a chromatographic setting selected from a group of settings including a loading solution and an elution condition.
[0018] In an embodiment according to the first aspect, the method further comprises applying the obtained chromatographic settings to a chromatographic separation process to separate the at least one target molecule.
[0019] In an embodiment according to the first aspect, wherein selecting an adsorption model comprises selecting a plurality of adsorption models from the set of predetermined adsorption models, wherein calculating a partition coefficient value comprises calculating a partition coefficient value for each of the plurality of adsorption models, wherein calculating an integral comprises calculating an integral for each of the plurality of adsorption models, calculating an aggregate distance metric for each of the plurality of adsorption models, selecting a new candidate value of an adsorption model parameter for each of the plurality of adsorption models, repeating the steps of calculating a partition coefficient value, calculating an integral, calculating an aggregate distance metric and selecting until the aggregate distance metric is minimized for each of the plurality of adsorption models, and determining the parameters of the adsorption model using the selected values of the adsorption model parameters associated with the adsorption model of the plurality of adsorption models that minimizes the aggregate distance metric (595).
[0020] At least one advantage of the disclosure according to the first aspect is that model parameters can be quickly derived from relatively few experiments conducted at low loading densities.
[0021] According to the second aspect of the present invention, the above-mentioned purpose and other purposes are achieved by a device, which includes a control unit, and the control unit includes: a circuit module, the circuit module includes a processing circuit module and a memory, the memory includes instructions executable by the processing circuit module, wherein the control unit is communicatively coupled to a controllable unit of a fluid network, whereby the device is configured to execute the method according to the first aspect when the instructions are executed by the processing circuit module.
[0022] According to the third aspect of the present invention, the above-mentioned purpose and other purposes are achieved by a computer program, which includes computer executable instructions, which are used to enable the control unit to perform any method steps according to the first aspect when the computer executable instructions are executed on the processing circuit module included in the control unit.
[0023] According to a fourth aspect of the present invention, the above objects and other objects are achieved by a computer program product, which includes a computer-readable storage medium having a computer program according to the third aspect implemented therein.
[0024] An advantage of the present disclosure according to the second to fourth aspects is that model parameters can be quickly derived from relatively few experiments performed at low load densities. By considering the following detailed description of one or more embodiments, those skilled in the art will gain a more complete understanding of embodiments of the present invention, as well as an appreciation of its additional advantages. It should be understood that similar reference numerals are used to identify similar elements shown in one or more of the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 Shown are graphs showing the properties of elution points measured in different operating modes in two or more chromatographic runs.
[0026] Figure 2 A computational grid, eluted measurement points, and surface are shown in accordance with one or more embodiments of the present disclosure.
[0027] Figure 3 An aggregate distance metric in accordance with one or more embodiments of the present disclosure is shown.
[0028] Figure 4 Paths in a computational grid are shown in accordance with one or more embodiments of the present disclosure.
[0029] Figure 5 A flow chart of a method according to one or more embodiments of the present disclosure is shown.
[0030] Figure 6 Shows about Figure 5A first embodiment of the method is described.
[0031] Figure 7 Shows about Figure 5 A second embodiment of the method is described.
[0032] Figure 8 Shows about Figure 6 and / or Figure 7 Embodiments of the methods described.
[0033] Fig. 9 An apparatus according to one or more embodiments of the present disclosure is shown.
[0034] By considering the following detailed description of one or more embodiments, those skilled in the art will gain a more complete understanding of embodiments of the present invention, as well as a recognition of additional advantages thereof.It should be understood that like reference numerals are used to identify like elements shown in one or more of the accompanying drawings. DETAILED DESCRIPTION
[0035] In order to predict or simulate any kind of chromatographic process, the governing natural principles must be described by mathematical model equations. In fact, column chromatography models usually consist of three individual but coupled models: a fluid dynamics model describing the general flow and transport patterns in the column, a model describing the intrapore concentration transport of the beads (if applicable), and an adsorption isotherm that models the adsorption process by thermodynamics. A large number of such models are known in the literature, which describe different chromatographic and interaction patterns. Given a model and the corresponding model parameters, the resulting chromatogram can be predicted by computer simulation.
[0036] Mechanistic models take into account physical and biochemical influences through natural laws. For example, outside the adsorber beads in a chromatography system, the injected components are transported by convection effects, which are usually caused by a connected pump. Dispersion effects include molecular diffusion and non-ideal flows, among other effects. Within the pores of the adsorber particles / beads, diffusion dominates the movement of the fluid. Finally, the component is adsorbed onto the inner surface of the particles / beads.
[0037] Examples of mechanistic models can be found on the following web pages: https: / / www.cytivalifesciences.com / en / us / solutions / bioprocessing / knowledge-center / what-are-mechanistic-chromatography-models .
[0038] Examples of isotherm models can be found on the following web pages: https: / / www.cytivalifesciences.com / en / us / solutions / bioprocessing / knowledge-center / adsorption-isotherm-models .
[0039] Examples of other models can be found at https: / / doi.org / 10.1016 / j.jbiotec.2007.05.036 as well as https: / / doi.org / 10.1016 / i.chroma.2021.462397 be found.
[0040] The "or" in this specification and the corresponding claims should be understood as a mathematical "or" covering "and" and "or", and should not be understood as "exclusive OR" (XOR, exclusive OR). The indefinite article "a" in this disclosure and claims is not limited to "one", and can also be understood as "one or more", that is, plural.
[0041] In this disclosure, references to processing circuitry modules and processing components will be interchangeable.
[0042] In the present disclosure, the term "target molecule" refers to a desired molecule, typically a molecule separated in a liquid chromatography process.
[0043] In the present disclosure, the term "chromatography" or "liquid chromatography" refers to a process for separating the components of a mixture. The mixture is dissolved in a first substance called a mobile phase. The mixture and the first substance are fed to the inlet of a chromatography column filled with a second substance called a stationary phase. The different components of the mixture travel through the stationary phase at different speeds, allowing them to be separated from each other at the outlet of the chromatography column. The characteristics of the specific mobile phase and stationary phase determine which substances travel faster or slower, and how well they are separated. These different travel times are called retention times and are recorded as retention data.
[0044] In the present disclosure, the term "chromatographic run" refers to a cycle in which the mixture and the first substance travel from the inlet to the outlet of the chromatography column.
[0045] In the present disclosure, the term "operating mode" defines the characteristics of the chromatography run. For example, the characteristics of the chromatography column, the characteristics of the beads / microwell volume, the characteristics of the mobile phase, the characteristics of the stationary phase, and the characteristics of the mixture. This may include mobile phase modifiers such as pH, salt concentration, and ionic strength, which may change during the chromatography run.
[0046] In the present disclosure, the terms "retention behavior" and "retention behavior data" refer to the behavior of the mixture and the first substance through the chromatographic column and the recorded data of the behavior. The retention behavior data may also include the chromatographic run characteristics defined above. The retention behavior data may also include processed or aggregated data based on the above data categories. In one example, the retention behavior data includes the travel time of the mixture components from the inlet of the chromatographic column to the outlet of the chromatographic column. In another example, the retention behavior data includes the running volume or total volume of the mixture and the mobile phase.
[0047] In the present disclosure, the term "elution point" refers to the peak of a mixture component that reaches the outlet of a chromatographic column and is detected at the outlet. In one example, an "elution point" refers to a peak in a chromatogram of a chromatographic run. Detection of the components that arrive can be performed using a spectrophotometer, a mass spectrometer, an ultraviolet (UV) detector, or fraction analysis. With respect to the recorded signal, an "elution point" can be defined, for example, as the point of the highest signal value, as the last maximum value of the signal, or as the centroid.
[0048] In the present disclosure, the term "properties of the measured elution point" or "properties of the calculated elution point" refers to the retention behavior data defining the measured elution point, for example, the travel time from injection at the inlet of the column to detection at the outlet of the column, the total volume of the mixture and the mobile phase composition at the time of elution, including the current ionic strength, salt concentration, current pH, current elution gradient slope, etc.
[0049] The injection point or starting point represents the chromatographic characteristics, e.g., the time point, when the components of the mixture reach the inlet of the chromatography column. In one example, the "point of injection" or POI represents the first point in the chromatogram from the chromatography run. In another example, the POI is defined as the time point when half of the sample volume has been injected.
[0050] In the present disclosure, the term "adsorption model" means a model that at least simulates the adsorption equilibrium. Typically, an adsorption model describes the concentration variation of the adsorbed mixture as a function of the concentration of the mixture in the mobile phase and the concentration bound to the adsorbent / beads, taking into account the properties of the material and the mobile phase. Examples of adsorption models are the Langmuir model or the colloidal particle adsorption (CPA) model. See also https: / / www.cytivalifesciences.com / en / us / solutions / hioprncessina / knnwledae-center / adsnrption-isotherm-modeis .
[0051] In the present disclosure, the term "partition coefficient value" means the ratio of the concentration of a substance in one phase and the concentration in a second phase when the two concentrations are in equilibrium.
[0052] In the present disclosure, the term "computational mesh" refers to a discretization of a computational domain consisting of elements, faces, and nodes. An example of a computational mesh is a Cartesian mesh. In a Cartesian mesh, the faces of the elements are aligned with the axes of a Cartesian coordinate system or a reference coordinate system. The nodes defined as the intersections of the faces can be placed evenly (equidistant) or unevenly. Therefore, an element (or cell) can be, for example, a line segment (1D), a rectangle (2D), or a cuboid (3D). Examples of computational meshes can be found in "Notes on computational Thermo-Fluid Dynamics)" DOI: 10.13140 / RG.2.1.5174.3124 The computational grid has an extent that effectively defines the maximum / minimum starting and ending points of the experiments or chromatographic runs used in the present disclosure.
[0053] In the present disclosure, the term "point in a computational grid" refers to a discrete point of a computational grid. Typically, each point in a computational grid includes as many (coordinate) values as the dimensions of the computational grid. That is, in a one-dimensional computational grid, each point includes a single value, in a two-dimensional computational grid, each point includes two values, and in an N-dimensional computational grid, each point includes N values.
[0054] In this disclosure, the term "path" or "path in a computational grid" refers to a set of values from a computational grid from a starting point (SP) to an end point (EP), such as points indicating changes in pH and salt concentration.
[0055] In this disclosure, the term "integral" or "computing integral" means a definite integral over a computation interval. An example of a method for computing a definite integral of a function defined on a computation grid over a given interval is the composite trapezoidal rule.
[0056] In the present disclosure, the term "aggregate distance metric" means a distance metric calculated using at least two partial distance metrics, e.g., partial distances between measured points of elution and properties of calculated points. An example of an aggregate distance metric is a metric used in the least squares method, where the distance measure is the residual sum of squares, which is minimized by varying candidate values of parameters. The residuals are typically calculated between the measured elution points and one or more properties of the calculated elution points.
[0057] In the present disclosure, the term "fluid network" refers to a fluid-coupled device component that is configured to hold, guide and control the flow of fluids. Examples of device components included in the fluid network are conduits / pipes, fluid containers / holders, sensors, valves, fluid columns, etc. Some device components may be controllable and communicatively coupled to a control unit and react to control signals received from the control unit and / or send data, such as device component characteristics and / or measurement data, as control signals for the control unit. Examples of such controllable device components are sensors and controllable valves. In short, the present disclosure eliminates the need for slow high-dimensional curve fitting for model calibration, particularly calibration isotherm adsorption models. The present disclosure also determines isotherm adsorption parameters within a linear range using only observed elution points, and does not require knowledge of the complete chromatogram. The advantage of the present disclosure is that model parameters are quickly derived from relatively few experiments conducted at low load densities.
[0058] The present disclosure identifies isotherm equation parameters that are valid in the linear range by applying a generalized method inspired by the Yamamoto method, but can be used for any given model equation, thereby enabling rapid identification of model parameters outside of LMA type models. In addition, an approximate calculation of the elution point is performed for any given model equation, thereby also supporting the use of more than one mobile phase modifier and nonlinear gradient elution experiments, which are known shortcomings of the traditional Yamamoto method.
[0059] At least one advantage of the present disclosure is that the computational complexity and time required to obtain a calibrated chromatographic model can be significantly reduced, and thus the liquid chromatograph setup can be significantly reduced.
[0060] Figure 1 A graph showing the properties of the measured elution points M-POE for different operating modes in two or more chromatographic runs is shown. Figure 1 As shown, different combinations of chromatographic run characteristics / operating modes result in different elution points. For example, the characteristics of the chromatographic column, the characteristics of the micropore volume / beads, the characteristics of the mobile phase, the characteristics of the stationary phase, and the characteristics of the mixture. The properties of the M-POE can be represented by the mobile phase modifiers 101, 102, such as the pH, salt concentration, and ionic strength at which the peak is detected.
[0061] In this example, a three-dimensional graph is shown. It should be understood that any number of dimensions may be used without departing from the present disclosure. pH is shown on the Y-axis, salt concentration is shown on the X-axis, and parameter GH is shown on the Z-axis. Parameter GH 103 may indicate normalized gradient slope, which is an operating characteristic used in conventional Yamamoto methods.
[0062] Figure 2 The calculation grid MTX, the surface S, and the measured elution point (M-POE) according to one or more embodiments of the present disclosure are shown. Figure 2 2 , two-dimensional computational grids 101 and 102 are shown. However, it should be understood that computational grids of any dimensions may be selected according to the application without departing from the disclosure of the present invention.
[0063] In one example, the computational grid MTX is arranged in one dimension, where the pH varies during a chromatographic run according to a specific gradient, and the elution points are characterized by the "normalized gradient slope" GH 103 and 102 pH at peak detection.
[0064] In a further example, the computational grid is arranged in one dimension, where the salt concentration C varies according to a specific gradient during a chromatographic run, and the elution point is characterized by the salt concentration at the peak detection and the "normalized gradient slope" GH.
[0065] In another example, the computational grid is arranged in two dimensions, where the pH value along with the salt concentration C is varied during the chromatography run. Figure 2 This is the case in the example shown. The elution point of the target molecule is shown to have three characteristic properties.
[0066] When an adsorption model predicts the outcome of a chromatography experiment, it can be used to generate predicted elution points, which naturally must lie on a hypersurface in the grid.
[0067] The surface S is generated by applying candidate values of the parameters to the adsorption model and then simulating / calculating the properties of the elution point for the breadth of the computational grid. In other words, when the pH and salt concentration C at elution are used as inputs to the model, the model can be used to calculate the normalized gradient slope that results in a given elution point characteristic of the surface. Different candidate values are used to minimize the distance between the measured elution point characteristic M-POE and the surface S / calculated using the model / partitioning coefficient values. Since multiple distances between the measured elution point property M-POE and the measured surface / property are taken into account, this effectively leads to minimizing the total distance D.
[0068] In one example, the computational grid is arranged in three dimensions, wherein the pH along with the first salt concentration C1 and the second salt concentration C2 varies according to a specific gradient.
[0069] This solution has the advantage that nonlinear surfaces can also be taken into account.
[0070] Figure 3 The aggregate distance metric D according to one or more embodiments disclosed herein is shown. The distance between the surface S and the corresponding measured elution point characteristic M-POE_i is calculated
[0071] In one embodiment, the aggregate distance metric D is a metric used in the least squares method, where the distance metric is the residual sum of squares minimized by varying the parameter candidate values.The residual is then formed by the distance from the surface S to the corresponding property M-POE of the measured elution point.
[0072] Exemplary properties of the measured elution point M-POE are the travel time from the inlet to the outlet of the column, the total volume of the mixture and the mobile phase, the current salt concentration, the current pH, the current gradient slope, etc.
[0073] Depending on the application, other curve fitting methods known in the art may be used. Examples of other curve fitting methods are Bayesian linear curve fitting or ridge regression.
[0074] Figure 4 Paths T1, T2 in a computational grid MTX according to one or more embodiments of the present disclosure are shown. These paths indicate how the operating mode changes over a series of chromatographic runs.
[0075] In one example, the retention behavior of a molecule in a first chromatographic run is described using a first path T1, where pH 102 is constant and the salt concentration C 103 varies according to a first gradient defined by the extent of the computational grid MTX. In other words, the salt concentration increases continuously from the starting point SP1 until elution at the end point EP1,
[0076] In another example, a second path T2 is used to describe the retention behavior of a molecule in a second chromatographic run, the salt concentration C is constant, and the pH varies according to a second gradient defined by the width of the computational grid MTX. In other words, from the starting point SP2, the pH value is continuously increased until elution at the end point EP2.
[0077] Figure 5 A flowchart of a method 500 according to one or more embodiments disclosed in the present invention is shown. In one or more embodiments, the method 500 is executed by a computer. The method includes:
[0078] Step 510: Performing an experiment as a series of chromatography runs. The series of chromatography runs includes at least two chromatography runs using different operating modes.
[0079] The mode of operation is further described in the opening part of the description.
[0080] In one example, in each chromatographic run of the series of chromatographic runs, the pH is varied according to a specific gradient. In another example, in each chromatographic run of the series of chromatographic runs, the pH and the salt concentration C are varied according to a specific gradient.
[0081] Step 520: Recording retention behavior data of at least one target molecule of each of the chromatographic runs, wherein the retention behavior data at least indicates a measured elution point M-POE. The measured elution point M-POE has the property M-POE i as defined above.
[0082] Retention behavior data may include any property or characteristic of a chromatography run. Examples of recorded retention behavior data are retention time or the time it takes a target molecule to travel from the inlet to the outlet of a column, run volume or the volume of mixture used in a chromatography run.
[0083] Step 530: Generate at least one calculation grid MTX in at least one dimension, wherein the width of the grid indicates the maximum range between the injection point / starting point SP and the elution point / end point EP.
[0084] The computational grid is further described at the beginning of the specification. In one example, the computational grid has two dimensions, with pH indicated in one dimension and salt concentration C indicated by a second dimension. In this example, the grid can be viewed as a grid where each intersection of the grid is represented by a pH and C value pair.
[0085] In other words, a two-dimensional computational grid MTX is generated for the breadth of the first pH range and the second salt concentration range C. The grid can then be formed using a suitable step size along the first and second ranges.
[0086] Step 540: Select an adsorption model from a set of predetermined adsorption models. The adsorption model has initial candidate values of adsorption model parameters.
[0087] The adsorption model is further described in the beginning of the instructions and on the following web pages: https: / / www.cytivalifesciences.com / en / us / solutions / bioprocessing / knowledge-center / adsorotion-isotherm-models .
[0088] Step 550: Use the selected adsorption model to calculate the partition coefficient values of all points in the calculation grid.
[0089] In one example, the partition coefficient value K is calculated according to the following relationship: Among them, k eq , k eq , 1, ν, ν1 are molecule-specific parameters of the model and Λ is a constant.
[0090] Step 560: Calculate the integral of the function of the calculated partition coefficient value on at least two paths in the calculation grid MTX to obtain the calculated elution point property C-POE_i.
[0091] Paths and computational grids are further described at the beginning of the description.
[0092] Step 570: Calculate the aggregate distance metric D from the calculated elution point property C-POE_i to the corresponding experimental reference property value M-POE_i of the measured elution point.
[0093] Combination Figure 3 The aggregation distance metric is further described.
[0094] Step 580: Select new candidate values for adsorption model parameters.
[0095] In one example, the new candidate value of the adsorption model parameter is selected from a predetermined set of candidate values. An example of a parameter of the adsorption model is k as in the above example. eq , k eq，1 ,ν,ν1.
[0096] Step 590: Repeat the steps of calculating partition coefficient values 550, calculating integrals 560, calculating aggregate distance metrics 570 and selecting 580 new candidate values until the aggregate distance metric D is minimized.
[0097] Step 595: Determine parameters of the adsorption model using selected values of the adsorption model parameters that minimize the aggregate distance metric.
[0098] In one example, the parameters of the adsorption model are determined by using the latest parameters of the adsorption model selected in step 580 .
[0099] In one embodiment, the measured elution point M-POE includes the selection of the time point during elution and / or the selection of the volume and / or the selection of the concentration of the mobile phase modifier.
[0100] Figure 6 Shows about Figure 5 A first embodiment of the method is described. In this embodiment, the breadth of the grid indicates the range of mobile phase modifier concentrations 101, 102 obtained during the chromatographic run. A path can be defined by any subset within the breadth of the computational grid MTX, i.e., from the injection point / starting point SP to the elution point / end point EP in one or more dimensions. The functions include inverse functions. Calculating the aggregate distance metric (D) includes:
[0101] Step 5701: Use the grid MTX to calculate the property C-POE_i of the elution point in the following manner: use the inverse function of the calculated partition coefficient value to calculate the integral on the two or more paths in the calculation grid MTX to obtain the property C-POEi of the elution point.
[0102] Step 5702: Calculate the aggregate distance metric D between the calculated one or more properties of the elution point C-POE_i and the measured one or more properties of the elution point M-POE_i.
[0103] Alternatively or additionally, the method further comprises:
[0104] Step 5703: Select a chromatographic column model from a set of predetermined chromatographic column models.
[0105] Step 5704: Determine candidate values for additional adsorption model parameters by fitting the retention behavior data to the simulated chromatograms. Examples of additional adsorption model parameters are parameters describing the binding kinetics, or parameters identifiable only in the nonlinear isotherm range that describe, for example, steric effects or molecule-molecule interactions.
[0106] Figure 7 Shows about Figure 5A second embodiment of the method described. In this embodiment, the extent of the computational grid MTX indicates the temporal duration of the series of chromatographic runs and the axial column size of the series of chromatographic runs. Calculating the aggregate distance metric D comprises:
[0107] Step 5706: Select a chromatographic column model from a set of predetermined chromatographic column models.
[0108] Alternatively or additionally, the method further comprises:
[0109] Step 5707: Using the selected chromatographic column model and the partition coefficient value calculated using the computational grid, calculate the propagation velocity value of the at least one target molecule from the column inlet to the column outlet.
[0110] Alternatively or additionally, the method further comprises:
[0111] Step 5708: Simulating the retention behavior data of the at least one target molecule by calculating the property C-POE_i of the elution point by calculating one or more integrals of the propagation velocity to obtain at least one trajectory passing through the calculation grid MTX.
[0112] Alternatively or additionally, the method further comprises:
[0113] Step 5709: Select candidate values for additional adsorption model parameters by fitting the retention behavior data to the simulated chromatogram.
[0114] Figure 8 Shows about Figure 6 and / or Figure 7 The method further comprises:
[0115] Step 810: Calibrate the mechanism model using the determined values of the adsorption model, the column model parameters and the selected candidate values of the additional adsorption model parameters.
[0116] Alternatively or additionally, the method further comprises:
[0117] Step 820: Simulate a chromatography run using the calibrated mechanistic model to obtain a chromatography setting selected from a set of settings including a loading solution and an elution condition.
[0118] Alternatively or additionally, the method further comprises:
[0119] Step 830: Applying the obtained chromatographic settings to a chromatographic separation process to separate the at least one target molecule.
[0120] Additionally or alternatively, the method further comprises performing model differentiation:
[0121] Wherein selecting 540 an adsorption model comprises selecting a plurality of adsorption models from said set of predetermined adsorption models.
[0122] Calculating 550 the partition coefficient value includes calculating the partition coefficient value of each adsorption model in the plurality of adsorption models.
[0123] The calculating integral 560 includes calculating the integral of each adsorption model in the plurality of adsorption models.
[0124] Wherein an aggregate distance measure D is calculated (570) for each of the plurality of adsorption models.
[0125] A new candidate value for an adsorption model parameter is selected 580 for each of the plurality of adsorption models.
[0126] Therein, parameters of the adsorption model are determined 595 using selected values of the adsorption model parameters associated with the adsorption model that minimizes the aggregate distance metric among the plurality of adsorption models.
[0127] Fig. 9 An apparatus 900 according to one or more embodiments disclosed herein is shown.
[0128] The apparatus 900 includes a control unit 960, which includes circuit modules, such as a processor and a memory. The memory may contain instructions executable by the processor, whereby the apparatus is operable to perform any steps or methods described herein.
[0129] Additionally or alternatively, the device 900 may further include a controllable fluid network, for example, including fluidically coupled pipes, fluidically coupled and controllable valves, and fluidically coupled and controllable pumps.
[0130] Additionally or alternatively, the device may further comprise a chromatography column 930, the chromatography column 930 further comprising at least one fluid inlet 931 and one fluid outlet 932. The fluid inlet 931 and one fluid outlet 932 are typically fluidically coupled to the controllable fluid network.
[0131] The apparatus 900 may further optionally include a fraction collection unit 950. The fraction collection unit 950 is typically fluidly coupled to a controllable fluid network.
[0132] Additionally or alternatively, the device also includes a display 980 .
[0133] Additionally or alternatively, the device further comprises an input device 970 .
[0134] In one embodiment, the device further comprises one or more sensors configured to measure the characteristics of the fluid in the device and / or fluid network. The control unit 960 is further coupled to each sensor in communication. The sensor can be configured to apply pressure to the fluid in the device and / or measure its pH and / or conductivity. Any suitable sensor for performing measurements on the fluid can be used.
[0135] The device 900 can be in the form of, for example, a chromatographic device, an electronic control unit, a server, an onboard computer, a fixed computing device, a laptop computer, a tablet computer, a handheld computer, a wrist-worn computer, a smart watch, a smart phone, or a smart TV. The device 900 may include a processing circuit module, which is communicatively coupled to a transceiver configured for wired or wireless communication. The device 900 may also include at least one optional antenna (not shown in the figure). The antenna may be coupled to the transceiver and configured to transmit and / or transmit and / or receive wired or wireless signals in a communication network (e.g., Wi-Fi, Bluetooth, 3G, 4G, 5G, etc.). In one example, the processing circuit module can be any one selected from a processing circuit module and / or a central processing unit and / or a processor module and / or a plurality of processors configured to cooperate with each other. In addition, the device 900 may also include a memory. The memory may, for example, include a hard disk RAM, a disk drive, a floppy disk drive, a flash drive, or other removable or fixed media drive or any other suitable memory known in the art. The memory may contain instructions executable by the processing circuit module to perform any steps or methods described herein. The processing circuit module may optionally be communicatively coupled to a transceiver, a memory, one or more sensors (e.g., a pH sensor, a conductivity sensor, and a pressure sensor), or a selection of any suitable type of sensor capable of measuring a property of the device and / or a fluid processed by the device. The device 900 may be configured to send / receive control signals directly to any of the above units or external nodes, or to send / receive control signals via a wired and / or wireless communication network.
[0136] The wired / wireless transceiver and / or the wired / wireless communication network adapter may be configured to send data values or parameters as signals to the processing circuit module and / or receive data values or parameters from the processing circuit module to or from other external nodes, such as a measured pH or conductivity or a generated buffer solution volume.
[0137] In one embodiment, the transceiver communicates with the external node directly or via a wireless communication network.
[0138] In one or more embodiments, the input device 970 is configured to receive input or indication from a user and send a user input signal indicating the user input or indication to the processing circuit module.
[0139] In one or more embodiments, display 980 is configured to receive display signals indicative of a rendered object, such as a textual or graphical user input object, from a processing circuit module and display the received signals as an object, such as a textual or graphical user input object.
[0140] In one embodiment, the display 980 is integrated with the user input device 970 and is configured to receive a display signal indicating a presented object, such as a text or graphical user input object, from the processing circuit module and display the received signal as an object, such as a text or graphical user input object, and / or is configured to receive input or indication from a user and send a user input signal indicating the user input or indication to the processing circuit module.
[0141] In other embodiments, the device 900 may further include and / or be coupled to one or more additional sensors (not shown in the figure), which are configured to receive and / or obtain and / or measure physical properties related to the device and / or the atmosphere around the device, and send one or more sensor signals indicating the physical properties of the device to the processing circuit module. For example, a temperature sensor that measures the temperature of the ambient air.
[0142] In an embodiment, the communication network communicates using wired or wireless communication technologies, which may include at least one of the following: local area network (LAN), metropolitan area network (MAN), global system for mobile networks (GSM), enhanced data GSM environment (EDGE), universal mobile telecommunications system, long term evolution, high speed downlink packet access (HSDPA), wideband code division multiple access (W-CDMA), code division multiple access (CDMA), time division multiple access (TDMA), Wi-Fi, Voice over Internet Protocol (VoIP), LTE Advanced, IEEE802.16m, WirelessMAN-Advanced, Evolved High Speed Packet Access (HSPA+), 3GPP Long Term Evolution (LTE), Mobile WiMAX (IEEE 802.16e), Ultra Mobile Broadband (UMB) (formerly Evolution-Data Optimized (EV-DO) Revision C), Fast Low-Latency Access Orthogonal Frequency Division Multiplexing with Seamless Handover (Flash-OFDM), High Capacity Space Division Multiple Access and Mobile Broadband Wireless Access (MBWA) (IEEE 802.20) systems, High Performance Radio Metropolitan Area Network (HIPERMAN), Wavelength Division Multiple Access (BDMA), Worldwide Interoperability for Microwave Access (Wi-MAX), and ultrasonic communications, but are not limited thereto.
[0143] In addition, those skilled in the art recognize that the device 900 may include necessary communication capabilities, for example in the form of functions, components, units, elements, etc., for performing the solution of the present invention. Examples of other such components, units, elements and functions are: processors, memories, buffers, control logic, encoders, decoders, rate matchers, derate matchers, mapping units, multipliers, decision units, selection units, switches, interleavers, deinterleavers, modulators, demodulators, inputs, outputs, antennas, amplifiers, receiver units, transmitter units, DSPs, MSDs, TCM encoders, TCM decoders, power supply units, feeding devices, communication interfaces, communication protocols, etc. They are appropriately arranged together for performing the solution of the present invention.
[0144] In particular, the processing circuit module and / or processing component disclosed in the present invention may include a processing circuit module, a processor module and a plurality of processors, central processing units (CPUs), processing units, processing circuits, processors, application specific integrated circuit modules (ASICs), microprocessors, field programmable gate arrays (FPGAs) or one or more instances of other processing logic that can interpret and execute instructions that are configured to cooperate with each other. Therefore, the expression "processing circuit module" and / or "processing component" may represent a processing circuit module including a plurality of processing circuits, such as, for example, any one, some or all of the above-mentioned processing circuit modules. The processing component may also perform data processing functions for inputting, outputting and processing data, including data buffering and device control functions, such as user interface control or the like. Exemplary Embodiments
[0145] In the exemplary embodiment, it is assumed that there is a unique mapping of the elution time of the target molecule to the elution volume / retention volume. Using the mapping, the elution volume can be determined from the elution time and vice versa; the case of the elution volume will not be considered further. Only in the case of flow interruption is a unique mapping impossible, but in this case, elution will not have to be expected.
[0146] Calculation of the properties of the elution point in the case of linear gradient elution .
[0147] In the case of a linear gradient, there is also a unique mapping of retention time to the concentration of the mobile phase modifier, which is varied over a predetermined gradient. If the modifier concentration at elution is known, the retention time can be derived from the modifier concentration.
[0148] Thus, the remaining question to be answered is how the modifier concentration at elution can be determined for a given operating mode, here characterized by a specific gradient slope.
[0149] The mathematical derivation of Yamamoto's method is based on the following equation Show that the integral of the partition coefficient Kp(C) and the partition coefficient for the target molecule from concentration CB at the start of the gradient to concentration CR at elution equals the normalized gradient slope GH. Note that some textbooks explicitly include the partition coefficient for the modifier species K', writing the denominator as (Kp(C)-K'), while others assume that it is included in Kp(C). The latter notation is used here.
[0150] Assume the general case where the above integral cannot be solved easily, for example, it is numerically evaluated along a discrete path of points {C0, C1, ..., Cn} where C0 = CB and Cn corresponds to at least the modifier concentration at the end of the gradient.
[0151] If the modifier concentration at elution is determined for a given GH, the piecewise integration and summing is continued until the subtotal is equal to or greater than GH. In the first case, the waypoint Ci provides the value sought; in the latter case, a more precise value between C(i-1) and Ci can be determined by interpolation.
[0152] If one wishes to determine the normalized gradient slope GH associated with a given modifier concentration at elution, then piecewise integration and summing is continued until the subtotal Ci is equal to or greater than CR. In the first case, the current sum provides the value sought; in the latter case, a more precise value between the last subtotal and the total can be determined by interpolation. Calculation of properties of elution points in general
[0153] In the case of isocratic elution, step elution or more complex modes of operation, a bijective mapping of retention time to one or more mobile phase modifier concentrations cannot be assumed. Since modifier concentrations are therefore not suitable for sensitive error measurements of possible distance functions, only retention time calculations are considered below.
[0154] First, a grid is created that spans the length of the column in space and at least the duration of the experiment in time. The modifier concentration is determined on the space / time grid. Not all grid points are relevant below, but for simplicity's sake, the method of efficiency optimization is not discussed further.
[0155] Next, the relevant mobile phase modifier concentrations at the grid points are determined. This can be done, for example, with a template-based finite difference implementation of the convection-diffusion model. These values are the basis for calculating the Kp value for each point in the space / time grid.
[0156] An exemplary equation for the propagation velocity of an adsorbable low concentration solute is given by: prop (t) = u(t) / (ε tot+(1-ε tot )K p ), where u(t) is the linear flow rate, εtot is the total porosity, and Kp is the partition coefficient for the target molecule, which is a function of the mobile phase modifier concentration (ionic strength, pH, etc.).
[0157] In order to obtain the trajectory of the target molecule through the space / time grid towards the exit of the column, the distance traveled per time interval is calculated by integrating the propagation velocity at the current position in the space-time grid. This integration is continued until the distance traveled is equal to or greater than the column length. In the first case, the current time point provides the sought value; in the latter case, a more precise value between the previous and current time points can be determined by interpolation.
[0158] If the last point in the time grid is reached before the distance traveled in space is equal to or greater than the column length, then elution has not yet occurred. As long as the current propagation velocity is positive, the remaining time to reach the exit can be estimated. Sample Application
[0159] Assuming from the above The example model is to be calibrated using linear gradient experiments, then for example 7 experiments can be used, such as Figure 2 Schematically shown in FIG: 3 experiments, each with a different salt gradient at 2 constant pH values, and a center point at a medium pH value with a medium gradient slope.
[0160] In the traditional Yamamoto method, one would proceed as follows. For each of the 3 results at constant pH, one can apply the classic Yamamoto method to obtain k eq ·exp(k eq，1 ·pH) and the values of ν+ν1·pH. Then, the various parameters k are determined by solving the equations eq ,k eq,1 ,v,v1.
[0161] Consider the following experimental results with a capacity of ∧ = 0.5m Normalized gradient slope GH[M] Salt concentration at retention [M] Relative pH 0.0024 0.118 0.3 0.0012 0.11 0.3 0.000792 0.105 0.3 0.000744 0.151 -0.3 0.0024 0.178 -0.3 0.001008 0.155 -0.3 0.0012 0.135 0 The above classical process will lead to k eq =7.8e-4, k eq,1 = -11.7, ν = 7.2, ν1 = -4.5. However, the predicted salt concentration at the retention of the center point would be 0.129 M, which is unacceptably far from the measured value of 0.135 M.
[0162] Using the new method, based on the integration of the function of the partition coefficients on the grid, different values can be obtained: eq =3.9e-3, keq，1 = -3.9, ν = 6.2, ν1 = -0.63, which leads to a much better overall fit for all points and a predicted salt concentration of 0.133 M at retention.
[0163] Finally, it should be understood that the present invention is not limited to the embodiments described above, but also relates to and incorporates all embodiments within the scope of the appended independent claims.
Claims
1. A method (500), include: performing (510) the experiment as a series of chromatographic runs, wherein the series of chromatographic runs comprises at least two chromatographic runs using different operating modes, recording (520) retention behavior data of at least one target molecule for each of the chromatographic runs, wherein the retention behavior data indicates at least a measured elution point (M-POE) and a property of the measured elution point (M-POE_i), generating (530) at least one computational grid (MTX) in at least one dimension, wherein the extent of the grid (MTX) indicates a maximum range between a start point of the experiment and an end point of the experiment, selecting (540) an adsorption model from a set of predetermined adsorption models, the adsorption model having initial candidate values for adsorption model parameters, The selected adsorption model is used to calculate the partition coefficient values of all points in the calculation grid (550), calculating an integral (560) of a function of the calculated partition coefficient values over at least two paths in the calculation grid (MTX) to obtain a property of the calculated elution point (C-POE_i), calculating (570) an aggregate distance metric (D) from the calculated properties of the elution points (C-POE_i) to the corresponding properties of the measured elution points (M-POE_i), Select (580) new candidate values for the adsorption model parameters, repeating (590) the steps of calculating (550) partition coefficient values, calculating integrals (560), calculating aggregate distance metrics (570) and selecting (580) new candidate values until the aggregate distance metric (D) is minimized, and Parameters of the adsorption model are determined (595) using selected values of the adsorption model parameters that minimize the aggregate distance metric (D).
2. The method (500) according to claim 1, wherein the measured elution point (M-POE) comprises selection of a time point during elution and / or selection of a volume and / or selection of a mobile phase modifier concentration and / or chromatographic run characteristics.
3. The method (500) of claim 1 or claim 2, wherein the extent (MTX) of the grid is indicative of a range of mobile phase modifier concentrations (101, 102), wherein the function of the calculated partition coefficient value is an inverse function, wherein the aggregate distance metric (D) is calculated include: The property of the calculated elution point (C-POE_i) is calculated (5701) by: The inverse function of the calculated partition coefficient value is used to calculate the value of the calculation grid (MTX) integrating over said at least two paths to obtain the property of said calculated elution point (C-POE_i), An aggregate distance metric (D) is calculated (5702) between the calculated property of the elution point (C-POE_i) and the corresponding property of the measured elution point (M-POE_i).
4. The method (500) according to claim 3, further comprising: include: select (5703) a column model from a set of predetermined column models, Candidate values for additional adsorption model parameters are determined (5704) by fitting the retention behavior data to the simulated chromatogram.
5. The method (500) according to any preceding claim, wherein the extent of the computational grid (MTX) indicates the duration in time of the series of chromatographic runs and the axial column size of the series of chromatographic runs, wherein the aggregate distance metric (D) is calculated include: The properties of the calculated elution points (C-POE_i) are calculated by: selecting (5706) a column model from a set of predetermined column models, calculating (5707) a value of propagation velocity of the at least one target molecule from an inlet of the column to an outlet of the column using the selected chromatographic column model and the partition coefficient value calculated on the computational grid (MTX), and The retention behavior data of the at least one target molecule is simulated (5708) by: The property of the calculated elution point (C-POE_i) is calculated by calculating the integral of the propagation velocity to obtain at least one trajectory through the computational grid (MTX), wherein the aggregate distance metric (D) is calculated between the property of the calculated elution point (C-POE_i) and the corresponding property of the measured elution point (M-POE_i).
6. The method (500) according to claim 5, further comprising: include: Candidate values for additional adsorption model parameters are selected (5709) by fitting the retention behavior data to the simulated chromatogram.
7. The method (500) according to any preceding claim, further comprising: include: calibrating (810) a mechanistic model using the determined values of the adsorption model parameters, the selected candidate values of the chromatographic column model parameters and additional adsorption model parameters, and A chromatography run is simulated (820) using the calibrated mechanistic model to obtain a chromatography setting selected from a set of settings including a loading solution and an elution condition.
8. The method (500) according to claim 7, further comprising: include: The obtained chromatographic settings are applied (830) to a chromatographic separation process to separate the at least one target molecule.
9. The method (500) according to any preceding claim, wherein selecting (540) the adsorption model comprises selecting a plurality of adsorption models from said set of predetermined adsorption models, Calculating (550) the partition coefficient value includes calculating the partition coefficient value of each adsorption model in the plurality of adsorption models. The calculating integral (560) includes calculating the integral of each adsorption model in the plurality of adsorption models. calculating (570) an aggregate distance metric (D) for each of the plurality of adsorption models, selecting (580) a new candidate value for an adsorption model parameter for each of the plurality of adsorption models, repeating (590) the steps of calculating (550) partition coefficient values, calculating one or more integrals (560), calculating an aggregate distance metric (570), and selecting (580) until the aggregate distance metric (D) is minimized for each of the plurality of adsorption models, and Parameters of the adsorption model are determined using selected values of adsorption model parameters associated with the adsorption model of the plurality of adsorption models that minimizes the aggregate distance metric (D) (595).
10. A device (900), the device include: A control unit (960) comprising a circuit module, the circuit module comprising: a processing circuit module, and A memory comprising instructions executable by the processing circuit module, wherein the control unit is communicatively coupled to a controllable unit of the fluid network, whereby the device is configured to perform a method (500) according to any of the preceding claims when the instructions are executed by the processing circuit module.
11. A computer program comprising computer executable instructions, for causing the control unit to perform any of the method steps of the method (500) according to any one of claims 1 to 9 when the computer executable instructions are executed on a processing circuit module included in the control unit.
12. A computer program product comprising a computer readable storage medium having the computer program according to claim 11 embodied therein.