Contract modeling: Outsource modeling work to GoSilico
GoSilico’s DSPX simulation software utilizes mechanistic models to build digital twins of your downstream process. Mechanistic models are based on the fundamental laws of physics and biochemistry, denoted in terms of mathematical language. Unknown model parameters must be estimated from experimental data. Numerical methods and computer simulation are then used to simulate your downstream purification application. This sounds complex and, to be honest, it is not foolproof. Implementing such a disruptive technology is a complex task and requires profound consideration. One key question is whether to build up the expertise in-house or to engage external experts.
DSPX is designed for the highest usability. It seamlessly integrates into your lab environment and is fully compatible with standard FPLC and filtration systems. Software training sessions, a comprehensive user guide and many tutorials allow you to become an expert quickly. You do not need any prior expertise to get started but, of course, your individual learning curve strongly depends on your prior expertise.
It is a major concern for us to enable quick and sustainable success. GoSilico’s contract modeling service allows you to outsource model building activities using DSPX. To get started quickly and to guarantee a smooth transition, it may even be helpful to start with a two-pronged approach: benefit from GoSilico’s longstanding model building expertise while training your own experts in-house.
Over the past few years, we have modeled a wide variety of industrial and academic downstream purification processes. We have modeled and simulated monoclonal antibodies, MAbs, VLPs, vaccines and many more besides in columns, on membranes, etc. From single-column batch processes to multicolumn continuous operations – we will also find the optimal model for your individual process and facilities!
In close cooperation with you, we perform a well-established model building process: Based on data gained via a few initial experimental calibration runs in your lab and with careful regard towards the specifics of your process, we select, calibrate and improve the optimal model for your specific needs. Unknown process parameters are estimated by inverse curve fitting. Careful sensitivity analyses and model validation loops are performed to make sure the final model is perfectly tailored to your project goal.
The outcome of a joint contract modeling project is typically the process model itself, along with a detailed model building report including methods, assumptions, process parameters and uncertainties.
You can then utilize your digital twin to….
- optimize your process in silico by replacing lab experiments with cheap and fast computer simulation
- generate deeper process understanding by interpreting physical model parameters
- perform in silico PC/PV studies to better understand product quality attributes and their relation to process parameters
- scale your process up or down or fit it to a facility
- implement a process monitoring and/or control strategy
and much more.
Every process is individual and modeling a process is, as well. Please contact us via the contact form below for a non-binding consultation to learn about the modeling possibilities concerning your process. Based on the information you provide in one or two telephone calls, we will draft an individual project plan outlining the project characteristics, the project goals as well as the estimated effort.
While we perform the model calibration, you are responsible for providing the necessary experiments. The number of experiments is dependent on the desired model complexity. For example, an in silico chromatography model typically requires about four to eight experiments to investigate protein-ligand interactions. Additionally, up to three experiments can be performed to characterize the column to increase the resulting model quality. Depending on your application, we also recommended you perform additional offline analytics for the different proteins of interest.