We use cookies to enhance your browsing experience and analyze site traffic. By continuing to use this site, you agree to our cookie policy. Privacy Policy
Leveraging large-scale human framework libraries and AI structure prediction, matching optimal framework regions from a wider sequence space, precisely maintaining CDR-framework interactions — no back-mutations needed, higher humanization, shorter timelines.
Validated Results
Why Is This Needed?
"Murine monoclonal antibodies face significant challenges in clinical applications from immunogenicity (HAMA reactions) and insufficient effector functions. Humanization eliminates heterologous features and enhances ADCC/CDC effector functions — a core strategy for improving therapeutic efficacy and safety."
Key Advantages
Traditional CDR grafting is limited to highly homologous frameworks, unable to screen from a larger space for the truly optimal acceptor framework.
Grafting often causes affinity decline due to CDR-framework incompatibility, requiring multiple rounds of back-mutation trial and error.
Extensive expression, purification, and screening are time-consuming, relying on experience-based mutation design.
Breaking through traditional similar-framework limitations, exploring greater sequence space via massive human framework libraries, providing diversified framework candidates for subsequent molecular optimization.
AI-predicted antibody structures accurately model CDR-framework interactions, which is key to maintaining affinity after humanization.
No back-mutations needed — typically only one round of expressing 10~20 antibodies yields candidates with affinity and expression comparable to parental, significantly reducing time and experimental costs.
Scores, clusters, and screens candidates based on humanization level, stability, and predicted biophysical properties, outputting optimized recommendations.
Validation
In a collaboration with an international pharmaceutical client, AI humanization was performed on 2 murine monoclonal antibodies Ab2 and Ab3. Based on AI prediction results and model scoring, 10 humanized antibodies were selected for recombinant expression to measure expression levels and affinity.
All antibodies showed expression levels comparable to or higher than parental; in one system, 3 antibodies had higher affinity than parental, others within 3-fold difference, with 1 antibody exceeding the patent-reported humanized antibody in affinity; humanization level reached 90%-95%.
Pipeline
Fully automated AI humanization workflow, from sequence submission to recommended results in one step.
Submit VH/VL sequences → Structure modeling & CDR identification → AI framework design → Generate 10,000–100,000 variants with scoring & screening → Output recommended sequences, Germline alignment reports & risk site scoring
Report
Below are partial screenshots from the antibody humanization service delivery report, showing the complete workflow from sequence analysis to humanization scheme evaluation.
Use Cases
Best suited for humanization optimization before downstream development of murine mAbs.
Traditional CDR grafting often leads to decreased affinity, frequently requiring multiple rounds of back-mutation. The platform leverages a large-scale human germline framework library and structural compatibility assessment to identify framework combinations better suited for the parent CDRs, eliminating the need for back-mutation.
Affinity retention depends on the 3D support relationship between CDRs and frameworks. The platform evaluates Vernier regions, CDR interfaces, and key framework sites to avoid disrupting CDR conformation after framework replacement.
Back-mutations usually mean multiple empirical trial-and-error rounds and extra experimental cost. If structure-driven framework matching directly yields candidates with affinity and expression close to the parental antibody, validation cycles can be significantly shortened.
Humanness percentage alone can miss stability, risk sites, and expression concerns. The platform combines humanness, structural stability, predicted biophysical properties, and clustering scores to output top candidates for downstream experimental selection.
Overcome immunogenicity barriers while maintaining affinity and achieving significant improvements in humanization levels.