GERMANY — Boehringer Ingelheim has entered into an agreement with IBM to harness the power of generative artificial intelligence (genAI) and foundation models for advancing therapeutic antibodies.
This collaboration entails Boehringer utilizing an IBM-developed pre-trained AI model, fine-tuned with proprietary data contributed by the German pharmaceutical company.
While the financial details remain undisclosed, Andrew Nixon, Boehringer’s global head of biotherapeutics discovery, expressed confidence in establishing an unprecedented platform for expediting antibody discovery, ultimately addressing the needs of patients facing high unmet medical requirements.
The partners will employ in-silico methods, leveraging IBM’s foundation model technologies to accelerate the antibody discovery process.
This involves utilizing disease-relevant target information, encompassing sequence, structure, and molecular profiles, to generate new human antibody sequences in-silico.
IBM’s biomedical foundation model taps into a diverse range of publicly available datasets, including protein-protein interactions and drug-target interactions, to create pre-trained models.
These models design antibody candidates for specific targets, which are then screened through AI-enhanced simulation to identify and refine the top binders.
Boehringer outlined the validation process, involving the production of candidates in “mini scales” for experimental assessment.
Results from laboratory experiments will feed back into the AI model, enhancing its accuracy through iterative feedback loops.
The primary objective is to input data on disease targets, allowing the AI model to simulate experiments and identify potential antibody molecules with a high likelihood of success. The selected candidates undergo laboratory testing, with the outcomes further refining the AI model.
IBM’s foundation model technology has demonstrated efficacy in generating biologics and small molecules with relevant target affinities.
This collaboration builds upon IBM’s established presence in drug discovery, exemplified by its partnership with Pfizer in 2016, granting access to IBM Watson cognitive computing capabilities.
Boehringer’s broader strategy involves establishing a digital ecosystem for streamlined drug discovery and development through strategic partnerships.
Earlier this year, IBM’s genAI model showcased its ability to predict physico-chemical properties of drug-like small molecules.
The collaboration with Moderna in April exemplifies IBM’s continued exploration of cutting-edge technologies, including quantum computing, to advance the development of mRNA medicines.