ELRIG – Drug Discovery 2026
Conference dates: October 14-16, 2026
Stand: G08
Location: London, UK
Speaking sessions
Title: HT-SPR Technology as a Data Engine for AI-Driven Drug Design
Presenter: Daniel Bedinger, Director of Application Science, Carterra
Date and Time: Wednesday, Oct. 14 | 11:50 – 12:20
Abstract:Carterra’s HT-SPR platforms set the bar for throughput and scalability in biologics discovery and now bring that same throughput to small molecule and fragment screening. By delivering high-resolution kinetics at high throughput, Carterra instruments generate the volume and quality of binding data that machine learning models require, turning SPR from a validation step into a training-data engine for AI-guided design. The “One on Many” platforms are the instrument of choice for binding studies in Lab-in-the-Loop antibody discovery, while the Carterra Vega enables rapid screening and characterization of large analyte libraries for compound library screening and large-scale SAR. This talk reviews HT-SPR applications and implementation strategies powering the next generation of AI-driven discovery and model-training workflows.
Poster session
Poster Title: High-Throughput SPR Screening and Kinetic Characterization Using Carterra Vega™ with AI Binding Prediction Comparison
Presenter: Jonathan Popplewell, PhD, Field Applications Scientist, Carterra
Abstract: Carterra Vega™ is a high-throughput surface plasmon resonance (HT-SPR) platform that enables 48-analyte parallel analysis through its multi-channel flow cell, supporting high-quality binding data collection for up to two targets plus an internal reference in each channel. In this study, small-molecule compounds from the Maybridge fragment library were screened against human carbonic anhydrase II (hCAII) and XII (hCAXII) using Carterra Vega, allowing a 384-well plate to be screened in approximately 35 minutes while providing kinetic context. With fast cycle times and an optional integrated plate-loading robot, Carterra Vega can screen more than 20,000 compounds per day. SPR screening results were compared with binding predictions generated using Boltz-2, and selected hits were subsequently characterized by titration experiments to determine kinetic and affinity parameters. The comparison showed agreement between predicted and experimentally observed binders for a subset of compounds, while also revealing additional binders that were not strongly predicted by the model. Further analysis suggests that AI-based binding predictions align well with known chemical patterns represented in available data, while SPR screening can reveal additional binders beyond those learned patterns. Together, these results demonstrate how HT-SPR can efficiently identify and characterize small-molecule binders while generating experimental datasets that complement and inform computational prediction approaches in early drug discovery.