Antibody humanization

Humanization is a crucial step for transforming antibodies candidates from non-human sources into suitable therapeutics. Traditional grafting methods, where CDRs are transferred from the parental molecule to human frameworks, are time-consuming, typically low throughput, and often reduce binding affinity. Other machine learning models like Biophi tend to fall short in efficiency and success rate in retaining binding affinity. Our generative deep learning approach to humanization eliminates the need for grafting and has shown superior results in retaining binding affinity compared to e.g. Biophi. Achieving higher success rates with fewer screened variants increases efficiency in both time and cost.

Talk to our experts about your humanization needs

Our generative deep learning model overcomes the drawbacks of both traditional grafting approaches, which are often time and resource-intensive, and public deep learning models like BioPhi, which typically fail to retain binding sufficiently (or at all). Building on our proprietary AIGX protein language models (PLM), a generative interface was trained to create humanized variants agnostic to the parental species used. With outstanding success rates per generated humanised variant, it brings significant time and cost savings.

Watch our scientific presentation to learn more
Humanization process shown through antibodies.

Our deep learning model has proven to retain the binding affinity of parental antibodies exceptionally well.

Express and screen fewer candidates while achieving higher humanization success rates.

By combining a high-throughput humanization setup with high success rates per humanized variant, enables the humanization of many candidates earlier in your research.

Adaptable by design, our model allows alteration of default settings with alternative restrictions when desired, ensuring it always aligns with your research objectives.

Genovac, a leading CRO in antibody discovery, expression, and characterization, and ENPICOM partnered to provide a fully integrated humanization service. By combining ENPICOM’s generative AI model with Genovac’s state-of-the-art lab facilities and expertise, we offer a comprehensive and cost-effective service from parental sequence to expressed and fully screened humanized antibodies. Our humanization model has been successfully validated, showcasing its exceptional performance compared to open-source alternatives like BioPhi. Detailed supporting data is available upon request.

  • Highly efficient at preserving binding affinity of parental antibodies
  • Enables humanization at any stage of development
  • Cost and time effective
  • Species agnostic
View the case study
Genovac logo

A deep learning tool and service for next-generation antibody humanization.

Watch now

Evaluate candidates for stability and manufacturability using state-of-the-art tools from Oxford University.

Learn more