DSPX
Your DownStream Process eXecuted in silico
DSPX - Your downstream process executed in silico
DSPX is a comprehensive software for simulating downstream processes in silico. DSPX relies on the simulation of mechanistic models, describing the underlying phenomena to explain and predict complex, emergent behavior seen during processes, enabling a holistic, mechanistic insight and therefore prediction of the entire downstream train. This simulation suite enables any computer to become a bespoke, digital, process development laboratory, tailored to a specific use through the modular design.
GoSilico - the world leader in downstream process simulation
DSPX is developed by GoSilico, the leading force in mechanistic modeling for downstream processing. With its chromatography simulation software ChromX, GoSilico has provided solutions for seven of the ten largest biopharma companies, as well as a wealth of experience supporting SME’s to derive process understanding. It is this familiarity of bringing powerful, complex methods from the realm of academia to industrial labs through training, consulting and developing intuitive software that has led to the development of DSPX.
ChromX is now a part DSPX
With the launch of DSPX, our broadly established chromatography simulation software ChromX is now available as a part of a bigger, more comprehensive DSP simulation suite. When designing DSPX, our vision was to develop ChromX into a straightforward and intuitively understandable simulation tool and from there roll out simulation to the entire downstream process chain. DSPX is therefore designed in a modular way, and all unit operations can easily be combined together. Usability was the other key driving force when designing DSPX. Thereby, you can design the simulation toolbox you need to simulate your specific downstream process application. To improve usability, DSPX comes with auxiliary toolboxes, which facilitate lab integration, collaboration within a team, compliance and much more.
Customer voices
Our group has used ChromX as a tool to support exploration of column operation scenarios, mass transport analyses and new isotherm development as part of our chromatographic bioseparations research. It’s convenient, it’s well-documented, it’s well-supported, and – most importantly for fundamental research – it’s flexible. ChromX allows us to focus on the separation science questions we want to answer without having to re-invent rigorous numerical algorithms for column chromatography simulations.Prof. Dr. Todd M. Przybycien
We believe in educating the next generation of engineers ready to tackle industry relevant challenges. ChromX is the perfect tool to achieve this goal in bioprocessing: a friendly user interface to foster student’s engagement, full access to all model equations to help reinforce their understanding, and complete flexibility to cover any case study in chromatography simulation and optimisation!Dr. Simone Dimartino
The reason we chose to do this modeling with ChromX is that we saw the true benefits of it. It’s not a showcase. It’s not just beautiful, futuristic technology – it’s practical, it’s applicable to process development, and the results matter.Dr. Yuyi Shen
Mechanistic chromatography modeling can be a powerful tool for process development. Not only does it accelerate development time by reducing the experimental workload, but it also significantly improves process understanding. I believe that chromatography modeling is gaining enough momentum in the industry that it will become a staple of downstream process development in coming years.Dr. Steven Benner
Over the past few decades, science has laid thefoundation for computer-guided process development. We've done our job. Now it's up to you to take advantage of the scientific achievements and go in silico.Prof. Dr. Jürgen Hubbuch
Based on our experience, mechanistic models contribute significantly in the generation of process understanding, leading to a reduction of experimental work, enabling process optimization and supporting process characterization. ChromX enables the setup of mechanistic models in an attractive short time frame without the need to spend time on how partial differential equations are solved. Once the model is set up, the user interface is friendly for end-user process developers.Dr. María Iglesias González
A streamlined workflow to set the standard for Good Modeling Practice
DSPX follows an intuitive navigation and model building workflow from project definition to application. The main navigation reflects this: In the section ‘project’, you can set the scope of your project, see a dashboard of project activities, collaborators and utilized data resources. In the section ‘components’, you can define all of your experimental real-world assets like system components, buffers and molecules.
Then you can specify the mechanistic model you want to utilize to simulate your process. Next, you can calibrate your model and identify unknown parameters from experimental data. Once calibrated, you’re ready to apply your model to simulate your actual downstream process, e.g. sample a parameter range, optimize a process or perform single experiments in silico.
Plug-and-play: Component-centric data structure
Columns, resins, molecules and even methods are treated as individual items. All information regarding such an item is stored as digital representations of the respective real world component. All of these items are interchangeable and can be added to a project and optimized. This allows information to be stored easily and reused, while enabling comprehensive development, the fast combination of process parts and simple process alterations.
Better together: Collaborative model development
The DSPX Collaboration toolbox enables you to work on a project while being connected within a team. Multiple users can contribute to the same project, while a version management system gives you a full overview of who made which changes when.
Shared databases which contain system and model components, e.g. sensors, molecules, resins or columns, further help to provide colleagues with accessible process knowledge. Such a database can be saved on a flash drive, a local folder or even within a cloud. However, a database does not have to be project specific: Databases can be shared throughout various projects and serve as a central knowledge base, containing all previously simulated system and model components. Project-independent databases thereby enable straightforward knowledge management across time, distance and amongst different teams.
User-friendly and intuitive: Interactive help and software tutorials
All DSPX features and functionalities are comprehensively described in the interactive DSPX Help and the DSPX User’s Guide. By clicking on the Help icon in the main menu of DSPX, the interactive help section is opened. DSPX will then show all the relevant information regarding the specific section of the navigation you are currently in. The help texts serve as a general guideline on what the user can do in this section. At the same time, they provide a short overview of the background. In case you are interested in more detailed information, the help texts directly refer to the respective chapters of this manual. Using the search button further assists you in finding related content in the manual. Additionally, detailed software tutorials lead you through typical DSPX applications step-by-step and enable a steep learning curve.
Sampling and optimization in DSPX: Let the algorithms work
DSPX provides the option of testing a multitude of combinations of different process conditions such as pH, load density and salt concentration with its sampling functionalities. Here, the user has the choice of different automatic sampling strategies. Besides automatically generating samples, DSPX can also import parameter sets from Excel™ files.
DSPX further provides advanced optimization algorithms specialized in maximizing yield, purity and other objectives by varying process parameters. Simply state the desired objective function and parameter ranges and DSPX 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, DSPX allows the use of any of these algorithms for both model calibration and application.
Latin hypercube
Latin hypercube sampling (LHS) provides an evenly distributed set of process conditions, optimized for distance between points. This strategy provides the best granularity of the sampling results given the number of sampled points.
Parameter sweep
Parameter sweep sampling is based on a regular multidimensional lattice of parameter sets. This option is helpful if the effect of a single parameter is to be studied under different process conditions.
Normal distribution
Gaussian normally distributed sampling is extremely useful for robustness studies. The user simply states the confidence intervals for several process conditions, for example the setpoint plus and minus the accuracy to which the parameter is adjustable, and DSPX provides a probabilistic representation of the resulting process outcomes such as yield and purity.
Try DSPX
If you are interested in using DSPX for commercial purposes, please fill in the contact form below. We’ll get in contact with you to plan the individual implementation in your organization. If you’d like more information about the DSPX license model, special features and pricing or if you’d like to receive a free trial version of DSPX, please don’t hesitate to contact us.