BioMed X and Servier Launch First European XSeed Labs to Advance AI-Driven Bispecific Antibody Design
Key Insights
BioMed X (search) and Servier have launched the first XSeed Labs in Europe at the Spartners incubator in Paris-Saclay, implementing a crowdsourcing model that embeds academic researchers within pharmaceutical R&D facilities.
The inaugural research project focuses on developing an AI-empowered platform for sterically guided design of bispecific antibodies, addressing key physical constraints that limit therapeutic effectiveness.
The initiative seeks to overcome steric hindrance, epitope accessibility, and spatial dynamics challenges that currently evade systematic design principles in bispecific antibody development.
BioMed X (search), an independent biomedical research institute based in Heidelberg, Germany, announced the launch of its first European XSeed Labs initiative in partnership with Servier. The collaboration will implement BioMed X's proven incubator model at Servier's research and development site in Paris-Saclay, hosted at the Spartners incubator facility operated by BioLabs (search).
The XSeed Labs model represents an innovative approach to pharmaceutical research, crowdsourcing talented researchers from leading academic institutions worldwide and embedding them within established R&D facilities of pharmaceutical companies. This marks the first such initiative in Europe, following the successful establishment of the first XSeed Labs in 2024 at Boehringer Ingelheim's R&D campus in Ridgefield, Connecticut.
AI-Driven Platform for Bispecific Antibody Design
The inaugural research project, titled "New AI-Empowered Platform for Sterically Guided Design of Bispecific Antibodies," addresses a critical challenge in cancer therapeutics. The project aims to develop an AI-driven platform that enables rational design of bispecific antibodies by accounting for steric and structural parameters of dual target engagement.
Bispecific antibodies represent an emerging and powerful therapeutic drug class for cancer therapy, demonstrated by the clinical approval of immune cell engagers (ICE) and the development of dual-targeting antibody-drug conjugates (ADC). However, their therapeutic success faces significant physical constraints that current design approaches struggle to address systematically.
The fundamental challenge lies in the requirement for both binding arms of a bispecific antibody to simultaneously engage their targets under physiological conditions. Steric hindrance, epitope accessibility, and spatial dynamics serve as key limiting factors that currently evade systematic design principles, creating an urgent need for innovative platforms that can rationalize and optimize bispecific antibody architecture.
Research Challenges and Objectives
The collaboration seeks proposals addressing four key technical challenges:
Machine Learning Integration: Design and training of machine learning models to integrate structural data, binding geometries, and cell surface topology into predictive frameworks.
Epitope Optimization: Mapping and prediction of epitope combinations that enable optimal immune synapse formation for ICE and internalization pathways for ADC applications.
Dynamic Modeling: Development of in-silico simulation tools for dynamic modeling of bispecific antibody engagement under physiological conditions.
Experimental Validation: Validation of predictive models through experimental systems and comprehensive structure-function analyses.
The initiative prioritizes interdisciplinary proposals combining expertise in antibody engineering, AI/machine learning, structural biology, and immuno-oncology. Projects demonstrating originality and translational potential will receive priority consideration.
Global Call for Research Proposals
BioMed X (search) and Servier are inviting research project proposals from academic scientists worldwide, with submissions due by August 31, 2025, through the BioMed X Career Space platform. The selected research team will join BioMed X's expanding global network of research groups operating at the interface between academia and industry.
"We are excited that jointly with our partner Servier, we could expand our successful XSeed Labs model to Europe," said Dr. Christian Tidona, Founder and CEO of BioMed X (search). "Following our first XSeed Labs in the US, this is a great validation of our new model, showing that early innovation by academic talents placed inside an established pharma R&D campus is indeed possible and can create significant value for a pharma company's early R&D pipeline."
Céline Triquel, Head of R&D Open Innovation at Servier, emphasized the collaborative potential: "We are glad to launch another open innovation model at Servier: We'll have soon the great opportunity to host an XSeed Labs team at Spartners. This is an incredible opportunity for Servier's R&D teams who will be in direct and close interaction with the XSeed Labs team. And finally, a chance for the patients as we invest in R&D to speed up our journey towards new therapeutic solutions in oncology."
Expanding Innovation Model
The XSeed Labs model implements BioMed X (search)'s established global crowdsourcing and local incubation processes within pharmaceutical R&D campuses. Outstanding research talents recruited from leading academic institutions worldwide are relocated and placed inside established commercial R&D ecosystems, providing access to industrial infrastructure while maintaining connection to BioMed X's global academia-style research network.
BioMed X (search) operates with sites in Heidelberg, Germany, New Haven, Connecticut, and XSeed Labs in Ridgefield, Connecticut, alongside a worldwide network of partner locations. The institute performs biomedical research and drug discovery across oncology, immunology, neuroscience, platform technologies, and artificial intelligence, with all projects supported by leading pharmaceutical companies and conducted by early-career scientists recruited globally.