Reveal Chromatography Overview

Why model chromatography?

Significant scientific and engineering advances over the last decade have improved our ability to model protein-based pharmaceutical processes. Based on these advances, the bio-pharmaceutical industry has begun to leverage mechanistic modeling of chromatography processes to increase process understanding and improve process design.

The predictive power and therefore utility of mechanistic modeling for chromatography cannot be overstated. A well calibrated model can be used to inform:

  • the most important process parameters to evaluate during process characterization
  • the identification of critical process parameters
  • operational ranges for process parameters
  • in-process acceptance criteria for critical quality attributes
  • pooling strategy, manufacturing deviation investigations, process optimization activities, edge-of-range failures, and raw material utilization
  • the impact of feed-stream variability on yield and purity.

Reveal Chromatography is leveraging many decades of industry experience and state of the art software technologies to make this modeling efficient, intuitive, and accessible.

Application’s guiding principles

Reveal Chromatography was designed with the following guiding principles in mind:

  1. Accessible: Though powerful solvers have been available for several years to model chromatography processes and to develop scientific tools, until that power is made available to process development scientists, and purification scientists in the lab, the industry won’t benefit from these advances. That’s why we have developed an easy-to-use software, with an intuitive interface, to make modeling accessible to anyone.
  2. By scientists for scientists: The team behind Reveal Chromatography is led by biopharmaceutical manufacturing industry leaders, and is composed of experts in protein purification, fluid dynamics, and scientific software development in order to bring both the best techniques and the best technologies to an organization as quickly as possible.
  3. Powerful: Reveal leverages state-of-the-art packages which make tackling this problem possible. That includes the Chromatography Analysis and Design Toolkit (CADET) as its core solver, developed at the Institute of Bio- and Geosciences 1 of Forschungszentrum Jülich (FZJ) under supervision of Dr. Eric von Lieres.
  4. Customizable: Nothing is more frustrating than an application that can do a fixed number of things, but prevents users from building custom tools and analyses to tackle what is unique about their workflows. We have chosen to architect Reveal Chromatography to be both fully scriptable, and highly customizable. By offering a Python scripting interface inside the application, exposing state-of-the-art technologies, and including user friendly packages such as Pandas and Matplotlib, users can customize Reveal to suit their needs.

Overview of Reveal Chromatography’s features

As of version 1.0.0, the following features are available in the Reveal Chromatography application:

  • Interactive experimental data loader, from Excel input file and AKTA files, as exported by the AKTA software (supports both .csv and ascii formats).
  • Interactive simulation builder, to simulate the chromatography processes assuming a protein model, and an operational setup. Reveal currently includes support for Multi-component Langmuir, pH-dependent Langmuir, Steric Mass Action (SMA) and pH-dependent SMA type binding models.
  • Ability to run the CADET solver on a (list of) simulation(s) with one click. Runner leverages all available power in machines by utilizing multiple CPUs to run simulations in parallel for maximum speed.
  • Experimental or simulated chromatogram visualization tool, to display UV absorption, pH or conductivity during the chromatography process.
  • Ability to define and run CADET on large grids of simulations, to explore the impact of a set of parameters (binding, transport, or operational) on the chromatogram and the performances of the chromatography process. This provides a powerful way to calibrate a model as well as simplify process optimizations. Monte-carlo based explorations are designed to support process characterizations to exploring the probabilities and origins for process failures.
  • Ability to automatically calibrate a set of parameters (binding, transport, or operational) to best fit observed experimental data. A visualization tool allows to customize the cost function that defines the difference between experimental and simulated data, and visualize interactively the impact of these customizations on the best model.
  • Endless ability to customize the application and automate recurring tasks with custom scripts (to be written in the popular Python scripting language). The application embeds a Python code editor with syntax highlighting, an interactive Python console for easy development, and extensive documentation to empower scripting-inclined users. And since 100% of the application is written in Python, 100% of it can be customized/automated!