A-Alpha Bio, a synthetic biology and machine learning company that measures and engineers protein-protein interactions, and Lawrence Livermore National Laboratory (LLNL), a federally funded research and development center, on March 1 announced a collaboration to accelerate therapeutic antibody discovery against emerging COVID-19 variants using synthetic biology and machine learning. To support the collaboration, A-Alpha Bio has been awarded a $1.1M Department of Defense sub-contract, which the company will use to generate quantitative and multidimensional binding data on antibody-antigen interactions for COVID-19 variants.
Despite early advances in machine learning (ML) guided rational drug design, developing accurate ML models remains a challenge due to the lack of abundant and high-quality data required to train and validate these models. A-Alpha Bio’s AlphaSeq platform represents the first multidimensional and quantitative approach for measuring protein-protein interactions. The AlphaSeq platform can simultaneously measure millions of interactions between antibodies and antigens or antigen mutants for high-throughput determination of properties like affinity, specificity, cross-reactivity and epitope, thereby generating valuable data to fuel ML model building and optimization.
As part of the collaboration, A-Alpha Bio will measure millions of protein-protein interactions between ML-generated antibody libraries and panels of coronavirus variants that LLNL will use to refine their ML models for predicting antibody sequences with improved affinity to variants of concern and potential future variants that may emerge. Multiple antibody libraries derived from existing therapeutics with weakened binding against the delta and omicron variants will be synthesized and cloned into the AlphaSeq platform. Binding strength will be measured at a library-on-library scale between each antibody mutant and a panel of COVID-19 mutants to elucidate the relationship between antibody sequence and variant binding profile using ML.
“The team at LLNL shares our belief that machine learning must be central to a rapid therapeutic response against evolving COVID-19 variants and that access to high quality and quantity data is critical for success,” said Dr. David Younger, co-founder and CEO of A-Alpha Bio. “This collaboration perfectly combines LLNL’s expertise in high performance computing, simulation and machine learning with A-Alpha Bio’s unique capability to generate quantitative, multi-dimensional protein-protein binding data at scale. We look forward to working together to build ML models, powered by AlphaSeq, that accelerate the discovery and optimization of therapies against emerging viral variants”
Since the emergence of the pandemic, LLNL has used science, engineering and technology to help the research and medical communities with tools and information to better understand COVID-19. LLNL has developed expertise and capabilities in integrating big data and predictive simulation capabilities to develop a new understanding of biological complexity and enable more precise predictions of health risk, accelerate development of countermeasures, develop treatment options and improve outcomes.
“A-Alpha Bio has developed a technology that can simultaneously and at scale measure antigen-antibody binding affinities, thus enabling the iterative design and testing of thousands of antigen and antibody variants," said Daniel Faissol, Principal Investigator for AI-Driven Development of Biologic Countermeasures at LLNL. “We are excited to leverage this platform through a collaboration with A-Alpha Bio. The data AlphaSeq can generate will enable us to develop much more powerful predictive models that form a key component of our computational antibody and antigen design platform, allowing us to better prepare for and, respond to, novel pathogens.”