← Back to Home

Bringing Experimental Results Back to AI: Hansa Partners with Cradle to Design Protein Drugs for Autoimmune Diseases

The collaboration will enable protein design to address activity, stability and manufacturing requirements simultaneously, while continuously learning from Hansa’s experimental data. Whether it can shorten development timelines still requires laboratory validation; the promoted 18-month saving is not a result of this collaboration.

By SURL BioNews

A protein that looks promising on a computer still faces many hurdles before becoming a drug: it must have sufficient activity, a stable structure and reliable manufacturing. A new collaboration between Sweden’s Hansa Biopharma and protein engineering platform company Cradle aims to incorporate these requirements into the design process, using data from each round of experiments to guide the next round of improvements.

The two companies announced the collaboration on October 1, 2026. Hansa will introduce the Cradle platform into its drug discovery pipeline to design, optimize and advance therapeutic candidates for severe autoimmune diseases. Sofia Järnum, Hansa’s Head of Research and Early Development, said that after evaluating several AI providers, the company selected Cradle for its ability to provide support starting with the identification of initial active molecules, combining multiparameter optimization with active learning from Hansa’s own experimental data.

Multiparameter optimization means assessing several potentially competing properties simultaneously when modifying a protein. Changes that increase activity may not also benefit stability or manufacturing. According to an interview with R&D World, Hansa will bring its existing data, disease knowledge and protein engineering experience into the platform. Researchers will first compare candidate designs, then make and test the molecules, feeding the results back into the next round of design.

Hansa scientists will continue to operate this workflow, while manufacturing, translational science and clinical teams track progress. In the interview, Järnum noted that the laboratory work of physically making and testing molecules remains a bottleneck. Even if computers can propose sequences more quickly, the properties of candidate molecules must still be measured to determine whether the designs work.

The collaboration announcement stated that Cradle customers had reported reductions in development timelines of up to 18 months. However, Cradle CEO Stef van Grieken explained to R&D World that this figure came from vaccine projects and that the extent of time savings varies by application. It is the company’s statement about other projects and cannot be treated as a result Hansa has already achieved, nor is it sufficient to estimate when a new drug might enter clinical trials.

Background

Hansa’s technological foundation includes enzymes that cleave immunoglobulin G (IgG); its imlifidase is already used in Europe in specific kidney transplant settings. This provides context for understanding the direction of its protein drug development, but the announcement did not specify which candidates would use Cradle or limit the collaboration to existing products. The benefits of the new platform must be validated separately in each development project.

Currently available public information does not describe the size of the datasets Hansa will contribute, the specific optimization metrics or experimental results comparing the platform with its existing workflow. There is also no evidence of clinical efficacy arising from this collaboration. The substantive progress represented by this partnership is the integration of AI design into a drug company’s experimental cycle. The next question is whether it can reduce unsuccessful attempts and produce molecules capable of passing subsequent safety and efficacy validation.

References

  1. Dealroom
  2. Hansa Biopharma
  3. Cradle
  4. R&D World