In a recent gathering, BioNTech, the German biotechnology firm renowned for its COVID-19 vaccine developed in partnership with Pfizer, showcased its increasing reliance on artificial intelligence (AI) as a cornerstone of its future aspirations. The event, held in collaboration with its subsidiary InstaDeep, marked the company's inaugural AI summit. During this event, BioNTech executives spent two hours highlighting how their proprietary technologies and infrastructure could secure an advantage in the swiftly advancing field.
A focal point of the summit was the unveiling of a novel AI model designed to engineer antibodies from scratch. This development adds to a growing repertoire of similar tools emerging from academic and commercial entities. Additionally, BioNTech demonstrated ways AI could be utilised for speedy analysis of tumour tissue slides, interpreting obscure genomes, and pinpointing cancer protein segments that can be targeted by the immune system.
BioNTech has invested significantly in AI by acquiring InstaDeep for approximately $440 million in early 2023. Yet, there remains a cloud of curiosity within the scientific community regarding which of BioNTech's array of cancer therapies bear the influence of InstaDeep’s technological advancements. Ryan Richardson, BioNTech's Chief Strategy Officer, commented on the integration, stating that while the company monitors this coordination internally, details remain undisclosed as part of a broader long-term strategy.
InstaDeep, employing around 370 staff, enjoys significant operational independence to explore fundamental challenges within AI. BioNTech appears satisfied with these results, with Richardson noting that, in some domains, InstaDeep’s AI models have surpassed those of BioNTech's decade-long efforts within just a couple of years. Despite not delving into specific details, the company sees this model as instrumental in attracting top-tier AI talent, which might otherwise gravitate towards tech giants like Google or OpenAI.
One of BioNTech’s pioneering applications of AI lies within its personalised mRNA-based cancer vaccines, already undergoing clinical testing. Since 2017, the design of these experimental vaccines has been fully automated, utilising algorithms to encode cancer genome fragments with no human intervention. While the company acknowledges room for improvement, details on InstaDeep’s role in further advancements remain under wraps.
At the summit, CEO of InstaDeep, Karim Beguir, revealed the development of a Bayesian Flow Network (BFN), a technological breakthrough in AI that he anticipates will be more advantageous for processing continuous, discrete, and heterogeneous scientific data. The BFN model acts as the foundation for InstaDeep’s new technology, specifically a protein-creation platform known as ProtBFN. Recently, the company published a preprint of this approach and introduced a variant of the model for antibody design, termed AbBFN.
The antibody design technology demonstrated its capability to generate new structures by optimising particular antibody components, such as the critical CDR-H3 loop integral to an antibody's ability to latch onto targets. Despite these advancements, Richardson expressed that BioNTech is less captivated by simply applying new AI models to traditional antibody technology, focusing instead on combining groundbreaking AI with unique therapeutic platforms like mRNA medicines.
BioNTech’s commitment to open science was underscored by its publication of AI models and making the code publicly accessible, aligning with the prevalent open-source culture in tech industries. However, Beguir noted constraints exist: the protein modelling software remains accessible only through non-commercial licences, preserving BioNTech's exclusivity in product development and commercialisation capacities.
This strategic move seeks to facilitate engagement and improvement through feedback from the scientific community while simultaneously solidifying BioNTech’s presence as a frontrunner in integrating AI with biotechnology. As the company continues to develop this intersection, the world will be keenly observing how these innovations translate into tangible medical advancements.
Source: Noah Wire Services