ChromX: Highest usability and seamless lab integration
ChromX software offers numerous functionalities and features to make chromatography modeling as simple and versatile as possible.
All ChromX features are comprehensively described in the ChromX Online Help and the ChromX User’s Guide. To get help quickly, the software also provides a help button which points out relevant sections in the user’s guide.
Detailed software tutorials further lead you through typical ChromX applications step-by-step and enable a steep learning curve.
The Import Wizard is a ChromX feature for managing the transfer of experimental data from UNICORN® to ChromX.
The ChromX Import Wizard reads in data directly from ÄKTA® systems and utilizes UNICORN®’s native output files in .res or .zip format or data exported to Excel® files. All information about system and column parameters as well as feed and buffer compositions can be specified and directly imported into ChromX. The Import Wizard extracts UNICORN®’s method information, such as the time stamp of feedstock load and elution or the flow rates, directly from the UNICORN® result files. Through this, time-consuming and manual tasks are automated to reduce the model set-up time and to avoid introducing errors.
The ChromX Model Selection Wizard facilitates the selection of respective column models, pore models and adsorption isotherm model. It guides the user through the different process options and asks for specifications, e.g.
- the type of flow (axial flow through a column or a membrane or radial flow through a cylinder ring)
- additional mixing chambers for pre- or post-processing
- micro porous structure of the adsorber
Based on the user’s choices, ChromX then automatically suggests the model, that best fits the process.
Most standard adsorption isotherms are implemented in ChromX by default: Steric Mass Action, Langmuir, mixed mode, and many more. For more specific model applications, individual isotherm models may be required. ChromX can easily be extended with own adsorption isotherm models, allowing customized model building using rapid and robust isotherm prototyping.
New isotherms are easily implemented in a C++ template, requiring only little expertise with programming and mathematics. Several examples showcase the usage of this template implementing for example the Langmuir isotherm, the Steric Mass Action model, and a pH-dependent isotherm. These examples can easily be extended to derive an individual isotherm model. With a little more mathematical complexity, completely new isotherms are only a few lines of code away. After compiling as a shared library (DLL), ChromX recognizes these isotherms in its “Plugins” folder automatically and incorporates them in its high-performance simulation core.
Upon process scale-up, the flow path of chromatography skids and columns becomes more complicated: External void volumes in the skid and column increase, bubble traps are introduced into the flow path, and valve and pump configurations are changed. Those modifications can significantly contribute to the overall peak shape.
To keep the model quality high and to reduce the risk associated with process transfer and scale-up, these extra column effects should be modeled explicitly. ChromX allows the characteristics of the flow path to be mimicked by a series of dispersed plug flow reactors (DPFR) and continuous stirred tank reactors (CSTR) before and after the column.
The ChromX Peak Finder facilitates model calibration for ion exchange chromatography by automatically finding and evaluating the peak of the chromatogram. It is based on an adaptation of Yamamoto’s method, in which the protein’s charge parameter and its adsorption equilibrium parameter are derived from the salt gradient slope and the salt elution concentration of an eluting protein. This method requires experiments with a protein load concentration in the linear range of the adsorption isotherm. The chromatograms of these so-called linear gradient elutions (LGE) are imported into the ChromX Peak Finder, enabling an instant deterministic calculation of the steric mass action models “charge” and “equilibrium” parameters. Subsequently, the remaining model parameters which cannot be derived from LGE data can be estimated in ChromX. Both, the ChromX Peak Finder and the subsequent parameter estimation can be used for heterogenous protein feedstocks, eliminating the need for purified protein samples.
ChromX provides the option of testing a multitude of combinations of different process conditions like pH, load density and salt concentration with its sampling module. Here, the user has the choice of different automatic sampling strategies. Besides automatically generating samples, ChromX can also import parameter sets from Excel® files.
In addition to manual process optimization, ChromX provides optimization algorithms specialized in maximizing yield, purity and other objectives by varying process parameters. Simply state the desired objective function and parameter ranges and ChromX does the rest. This is realized with the help of several different heuristic and deterministic optimization algorithms, including gradient descent, adaptive simulated annealing and genetic algorithms. Due to its flexible nature, ChromX allows the use of any of these algorithms for both model calibration and process optimization.
ChromX is a powerful tool for standardized workflows and coherent data management. To enable a seamless integration of ChromX into an existing software tool chain and data handling workflows, ChromX comes with several options for third-party software integration, data import and export.
Naturally, all simulated chromatograms, optimized processes and sampling results can be visualized in ChromX. Data export to Excel® in .xls or .xlsx format is also possible and chromatograms and result plots can be exported in .jpg, .png, or .svg format.
Although mechanistic chromatography models are, in general, more sophisticated than statistical models, it might be feasible to continue using existing workflows for data analysis from the DoE software tool chain. Therefore, ChromX offers a bi-directional interface to statistical software such as JMP® or SIMCA®. Simulation results from ChromX can be exported to statistical software for visualization or data analysis. To support the transition from DoE workflows to mechanistic simulations, experimental designs created in JMP®, SIMCA® or Excel® can be imported to ChromX’s sampling module. In this way, experiments can still be designed the traditional DoE way but, instead of running lab experiments, ChromX is used to simulate the results. Keep in mind that, contrary to DoE-based models, which can only identify and predict trends and patterns within the experimental data, the simulation data of ChromX enables deeper mechanistic process understanding and the capability to extrapolate outside the calibrated parameter ranges.
Efficiency, data integrity and automation are key aspects of data handling and successful digital workflows. Therefore, ChromX can be run from command line and from a MATLAB® interface, relying on an xml-based data connection and a command line execution. In the opposite direction, it is possible to export simulated chromatograms and concentration profiles inside the column in .mat format for fast integration into MATLAB®.
As an alternative to visualization in MATLAB®, ChromX can export simulated chromatograms and concentration profiles inside the column as .vtk files, which can be used in sophisticated visualization tools such as ParaView®.