Location:
Classroom 3, Sanderson Building, King's Buildings, University of Edinburgh
Date:
All-Atom Modelling of Methacrylate-Based Chromatography Resins for Isotherm Parameter Prediction of Biomolecules
Abstract
Chromatographic separations rely on highly specific interactions between target molecules and chromatography resins. To elucidate these interactions, molecular simulations have proven as a powerful tool, shedding light on the adsorption mechanisms of biomolecules and providing predictive accuracy of their affinities for resin surfaces. Despite notable successes of previous studies in predicting the binding and elution behavior of proteins, the precise molecular modeling of chromatography resins poses a significant challenge, primarily due to the lack of detailed knowledge concerning the resins' all-atom three-dimensional structures. To address this challenge, our research presents a novel workflow for the generation of all-atom models of methacrylate-based chromatography resins. A subsequent workflow then utilizes these resin models to calculate the binding poses and associated binding energies of various target molecules in a high-throughput manner, currently accommodating entities up to 500 atoms in size. Both workflows incorporate multiple simulation steps and are fully automated to streamline the operational process.
As an initial validation of our approach, we applied these workflows to create an all-atom model of a commercially available multimodal resin produced by Tosoh Bioscience and proceeded to compute the binding energies of linear peptides with a wide spectrum of molecular weights. Subsequently, the model was trained with experimental data to generate a digital twin. The developed digital twin demonstrated a good predictive capability for experimentally determined Langmuir constants, achieving an R² of 0.94. In comparison with conventional simplified models, which assume a direct attachment of ligands to a flat surface, our all-atom model showed a marked improvement, reducing the root mean square error (RMSE) from 4.18 to 0.20 L/mmol. These advancements can not only help to facilitate a more nuanced understanding of biomolecular interactions with chromatographic media but also enhance the predictive modeling capabilities essential for the design and optimization of chromatographic separation strategies. Future studies will aim to expand the workflow’s target molecule size capacity to include proteins of various sizes.
Speaker
Tim Ballweg, currently a doctoral researcher at the Institute of Functional Interfaces, Karlsruhe Institute of Technology (KIT), has been contributing to the institute since 2021. His research primarily revolves around creating digital twins for adsorbents utilized in biotechnology, specifically concentrating on methacrylate-based chromatography resins and continuous flow catalysis with enzymes. Before embarking on this research journey, Tim gained practical experience through an internship at Tosoh Bioscience in Darmstadt in 2018. Here, he gained skills in high-throughput chromatography for process development.
Academically, Tim holds a Master of Science in Biochemical Engineering from the Karlsruhe Institute of Technology, graduating with distinction in 2020. His master's thesis was centered on the assembly and simulation-based process design of a micro-SMB.
Throughout his academic journey, Tim has consistently engaged in research and development in the field of biochemical engineering, with a particular focus on high throughput experiments and modeling in chromatography.
