Mary Vernet, Alliance Manager at Ailux Biologics, presented ‘Cutting Through the Hype: Four Practical Cases of AI Empowering Antibody Discovery,’ which aims to assess the real impact of AI in antibody discovery by presenting four practical case studies.
She said: “Everybody's talking about AI these days. And people are making a lot of different claims about what AI can do for us in antibody discovery. At Ailux, we wanted to ask the question, is AI really having an impact or is it a lot of hype?”
Vernet discusses four specific cases where AI has been incorporated into antibody discovery workflows at Ailux Biologics. These cases include antigen design, hit generation, and lead optimization and engineering.
The first case involves designing a new panel of antibodies for a challenging GPCR target using the XrossXeven platform, an AI-guided GPCR antigen design with large-scale mutation.
Vernet said: “In the end, we end up with AI design GPCRs with anywhere from 120 to 160 point mutations, so quite large scale mutations again, only in the TM and the intracellular region.”
This approach significantly improved the expression and stability of the GPCR mutants compared to the wild type.
In the second case, Vernet focuses on generating a new panel of hits with varying levels of functional activity using the XploreSeq platform, which combines NGS sequence analysis with AI for hit prediction. This method resulted in a high hit rate and a diverse panel of hits with favourable developability profiles7.
“The whole process can take a minimum of six weeks,” attested Vernet.
She continued: “We're able to identify diverse panels of hits because we're looking at the sequence data upfront. And in parallel with hit prediction, we can also select for favourable developability. Overall, this allows us to achieve a high hit rate.”
Vernet’s third case involves improving the affinity of a parental VHH antibody using AI-guided affinity maturation. The process, which included structural modelling and generative AI, resulted in a hundredfold increase in affinity within a short time frame.
“In the end we were able to identify our top clone which had an affinity of 1.53 nanomolar with just five point mutations and that resulted in a hundredfold increase in affinity.”
