This webinar will be hosted by The Antibody Society and presented by Dr. Jesper Sørensen, Head of Biomodeling at OpenEye, Cadence Molecular Sciences. Jesper received his Ph.D. in Nanoscience (chemistry) from Aarhus University in Denmark and subsequently did a post-doc in Prof. Rommie Amaro’s lab at UC San Diego in California. From there he joined a biotech company, Dart Neurosciences, as a scientific software developer. In 2018 Jesper joined OpenEye Scientific Software, where he co-developed the Spruce toolkit for protein structure preparation. He now manages a developer team focused on making advancements in drug discovery tools for biologics.
About this session
OpenEye’s suite of antibody modeling tools, accessed via the cloud platform Orion®, includes Specifica’s state-of-the-art antibody sequence analysis tool (AbXtract) and a set of open-source tools.
In this webinar, Dr. Jesper Sørensen will guide viewers through the process of modeling antibody structure on Orion, beginning with sequences from a selection campaign analyzed using AbXtract, or from internal or public databases. Full 3D structures are then generated with the AI-driven structure predictor ImmuneBuilder (requiring only around 100 ms per structure when predicting at the multi-thousand scale). From these structures, a wide variety of physico-chemical properties are calculated to enable the selection of those antibodies most likely to be successfully developed into human therapies. Structural diversity in the complementarity-determining region (CDR) can be further explored using knowledge-based loop modeling, while short trajectory and enhanced sampling molecular dynamics are used to explore global conformational diversity.
Structural families within computed conformational ensembles can be identified by clustering using 3D similarity based on a physically rigorous shape and chemical feature distribution within the CDRs. Individual sequences can be further optimized by single or multi-point mutations and/or loop replacement, and these new hypotheses can be submitted to further rounds of physico-chemical property calculation and conformational exploration, allowing rapid profiling of a wide variety of candidate antibodies.
