CAMBRIDGE, Mass.–Biointron Biological USA has reported laboratory validation results for artificial intelligence-designed antibodies developed using Molence Bio’s Mol-1 platform, demonstrating the potential for combining AI-driven drug discovery with experimental testing.
The company said 23 of 30 AI-designed antibody candidates met at least three of four criteria used to assess their suitability for further drug development.
The testing was conducted through Biointron’s RushDATA platform, a laboratory validation service designed to evaluate antibodies generated by artificial intelligence models.
Mol-1, developed by Molence Bio, is an AI model designed to predict antibody-antigen structures and generate functional nanobodies from scratch.
According to the companies, Mol-1 demonstrated performance comparable to IsoDDE, a leading closed-source model used in antibody-antigen structure prediction.
Biointron’s role in the project was to evaluate whether the AI-generated antibody designs possessed characteristics that could support their development into potential therapeutic candidates.
The company tested Molence Bio’s VHH-Fc antibody candidates using a series of laboratory assessments focused on properties associated with manufacturability, stability, and safety.
These evaluations measured four characteristics: hydrophobicity, thermal stability, self-interaction, and nonspecific binding.
The results showed that approximately 77% of the tested candidates met at least three of the four development criteria.
Biointron compared the experimental results against reference antibodies, including infliximab, bevacizumab, bococizumab, gantenerumab, and trastuzumab.
These comparisons allowed researchers to evaluate the AI-designed candidates against antibodies with established clinical or developmental profiles.
The findings represent an early step in determining whether computationally designed antibodies possess properties suitable for further pharmaceutical development.
However, meeting laboratory development criteria does not establish that an antibody will be safe or effective in humans.
Biointron said its RushDATA platform is designed to address a growing challenge in AI-driven drug discovery: translating large numbers of computationally generated antibody designs into experimentally validated candidates.
The platform combines high-throughput antibody production with laboratory testing to generate data that can be incorporated into subsequent AI modeling and optimization.
According to the company, the service has the capacity to process more than 10,000 antibodies per day.
Biointron said it can deliver experimental results in as little as two days using cell-free protein expression or three days using Chinese hamster ovary cell expression systems.
The platform also provides binding affinity measurements and assessments of antibody stability and molecular interactions.
Experimental findings are organized into structured datasets that can be accessed through application programming interfaces, allowing researchers to incorporate laboratory results into AI-based drug discovery workflows.
This approach is intended to create a feedback system in which computational predictions are tested experimentally, with the resulting data used to refine future antibody designs.
The growing use of artificial intelligence in biotechnology has enabled researchers to generate large numbers of potential therapeutic molecules. However, laboratory validation remains necessary to determine whether those molecules possess the characteristics required for further development.
The Mol-1 project illustrates how computational antibody design and laboratory testing can be integrated to identify candidates with potentially favorable development properties.
Biointron, founded in 2012, provides antibody discovery, development, and production services to pharmaceutical and biotechnology companies.
The company said it serves more than 3,000 organizations across more than 20 countries and regions, including major pharmaceutical manufacturers.
Biointron also provides antibody production services and cell-line technologies used in drug development programs, including projects that have advanced into international clinical trials.
The company said the Mol-1 collaboration demonstrates the potential for using experimental data to improve AI-generated antibody designs and accelerate early-stage therapeutic research.


