Downstream processing

By capturing the true physics and chemistry of chromatography, our tools predict separation performance, impurity clearance, and help you find optimal operating conditions. This enables you to design robust, high‑yield purifications with fewer experiments.

Ionic can predict the impact of numerous process parameters including (but not limited to): product concentration in the crude solution, impurity concentration in the crude solution, loading volume, flow rates, salt concentration, salt type, buffer concentration, buffer type, gradient slope.

Absolutely! Unlike statistical models which are based on correlations and are only valid for their calibration data set, mechanistic model parameters have an actual physical meaning and can be applied to different process setups and process scales. This gives you confidence in scaling up your chromatography processes by changing column dimensions and even enables you to adapt the operating sequence if needed. Mechanistic modeling can also help with troubleshooting. If the scaled-up process isn’t performing as expected, you can test potential reasons via simulation and take appropriate action based on the results. 

Yes. Built-in models for ion exchange, hydrophobic interaction, and reverse phase chromatography, which are commonly used for TIDES purification, are standard in Ionic. Other chromatography modes like normal phase, affinity and even complexing resins can also be simulated.

Usually, one particular mode of interaction between the target molecule and the stationary phase is dominant and it is sufficient to model that. Also, the built-in models generally have a sufficiently large scope to cover most purifications. Should this not be the case, we have the possibility to create user models considering specific and multiple (e.g. mixed mode) interactions.

Definitively! We offer several levels of support, ranging from essential to comprehensive, where process experts perform the modeling work for you so that at the end you have a ready-to-use model in your hands.

Often only a small set of targeted experiments is needed to parameterize a first reliable model. We guide you in collecting the minimal dataset required to achieve actionable results quickly. The exact number depends on the targeted precision, but typically around 15 experiments, of which only a handful are full purification experiments, are sufficient to build a model. Read more about our proposed methodology in an OPRD article (link to https://www.ypsofacto.com/publication-post/27-Oligonucleotide-Purification-by-Ion-Exchange-Chromatography-Step-by-Step-Guide-to-Process-Understanding-Modeling-and-Simulation).

While it is true that you need to perform some experiments for each new target molecule to feed data to the model, the more you do mechanistic modeling, the better it gets. Actually, a platform approach is particularly suited to mechanistic modeling, as some of the chromatographic media, buffer solutions and equipment pieces will likely be used for several target molecules. Only part of the overall experiments involve the use of the actual target molecule. Characterization of the solid phase, buffer solutions and equipment pieces only needs to be done once and can be re-used for another target molecule. Only those experiments needed to characterize the specific target molecule in solution and its interactions with the solid phase are required to be performed each time you have a new molecule.

Yes. Any kind of multicolumn chromatography process with any number of columns and configuration (e.g. MCSGP) can be simulated to help you increase productivity, reduce resin costs, and shorten cycle times before making hardware changes.

Yes. We support data import from common chromatography equipment and offer interoperability options that fit seamlessly into your digital workflow. If a direct import from the equipment is not possible, csv or Excel import can be used or a custom import could be tailored to your specific equipment.

It strongly depends on the complexity of the model and of the simulated process. Generally, most simulations are done in a range of a few seconds to a few minutes on common office laptops. Very complex and demanding processes may take a few hours to compute, but this is rare and still faster and less expensive than setting up, running and analyzing an experiment.

Yes. We will evaluate with you your actual collaboration needs and work with your IT department how to best integrate our software into your existing IT infrastructure.

Yes. We will evaluate with you your actual needs and work with your IT department how to best integrate our software into your existing IT infrastructure. Integration may be achieved through several ways: for example, by using standard data exchange formats like txt, csv, Word and Excel or by tailoring custom imports/exports to communicate with the other apps of your workflow.

Clients have reported up to four times fewer experiments, more than fifty percent solvent reductions, and doubled productivity in certain chromatography steps. These improvements lead to faster development, lower costs, and greener processes.

Yes. Mechanistic models are based on first principles and offer full transparency and explainability. Their assumptions, equations, and parameters are explicitly defined, allowing for established validation procedures. For a given input and set of model parameters, such models will always reliably provide the same output. This makes them suitable for regulatory purposes. The level of details regarding the model description in the dossier is dependent on the intended use of the model, its role in the control strategy and the risk to material quality.