Galux, a biotechnology company based in South Korea that focuses on AI guided protein therapeutic design, has reported new experimental results indicating that its computational platform, GaluxDesign, can produce high affinity antibody candidates from relatively small design sets. The company states that these findings illustrate a potential shift from probabilistic antibody discovery methods toward predictive and rational molecular design.
In the study, researchers generated fifty AI designed antibody sequences for each of eight independent epitopes. Across this set, the overall hit rate was 31.5 percent. Approximately 10.5 percent of all designs demonstrated affinity within a therapeutically relevant range, and several candidates exhibited picomolar binding strength. All candidates were evaluated in full length IgG format, which confirmed that the designed sequences retained properties compatible with therapeutic development without requiring extensive downstream engineering.
This performance represents more than a numerical improvement in success rate, it reflects a fundamental shift from discovering antibodies to rationally designing them. Traditional antibody discovery depends on probabilistic enrichment from massive libraries and requires months of iterative affinity maturation and humanization. The ability of GaluxDesign to deliver potent binders within weeks, and without heavy downstream engineering, demonstrates a true transition from stochastic discovery to rational molecular design.
Chaok Seok, CEO of Galux
Company representatives noted that the ability to obtain potent binders within a short design cycle contrasts with traditional discovery strategies that rely on large screening libraries, multi step enrichment processes and subsequent affinity maturation and humanization. The company’s leadership characterized the results as evidence that antibody design can be approached through deterministic modeling rather than stochastic selection.
The study follows earlier work in which Galux reported de novo generation of antibodies for eight distinct therapeutic targets, including PD L1, HER2, EGFR S468R, ACVR2A and ACVR2B, FZD7, ALK7, CD98hc and IL 11. That work yielded candidates with binding affinities as low as 9 pM. The company also resolved the structure of a designed PD L1 antibody bound to its target using cryo electron microscopy with a 1.1 angstrom interface root mean square deviation, demonstrating structural accuracy. Functional attributes such as subtype selectivity and mutant specificity were shown to be preserved when expressed in full length IgG form. According to the company, these earlier findings, combined with the present small set design results, indicate that GaluxDesign can repeatedly generate structurally faithful and tunable antibodies from limited computational input.
What these studies collectively show is that the essential qualities of a therapeutic antibody can all be defined at the design stage and realized experimentally through our platform. Our next step is to take this level of design control toward more advanced modalities, including multi-target binders and new functional architectures, and demonstrate how AI can expand the boundaries of what therapeutic antibodies can do.
Chaok Seok, CEO of Galux
Galux states that its platform integrates structural accuracy, functional developability and target diversity, and now demonstrates efficiency at small design scales. The company suggests that such capabilities may support a more predictable approach to therapeutic antibody discovery and development.
Future work will involve applying the design framework to more complex modalities, such as multi target binders and novel functional architectures. The company aims to assess whether algorithmic design can broaden the range of biological mechanisms that engineered antibodies can achieve.