[E-Daily Reporter Kwon Oh Seok ] #PROTEINA CO., LTD., an artificial intelligence (AI) drug discovery company, is set to fully launch the commercialization of its next-generation de novo antibody drug design business. The company announced on the 14th that, to this end, it will invest approximately 3.4 billion won in ABX Biosciences (hereinafter “ABX”) to become its largest shareholder. With this investment, PROTEINA CO., LTD. plans to expand its business scope as an AI-driven drug discovery company capable of handling everything from AI-based de novo antibody design to large-scale validation, and to establish the de novo antibody design business as a new growth engine.

ABX is the entity responsible for commercializing the AI antibody design platform “AbGPT-3D (tentative name),” which is currently under development as part of the Ministry of Science and ICT’s national project to “Establish a Cloud-Based AI Antibody Bank.” Professor Baek Min-kyung of Seoul National University, who is leading the development of AbGPT-3D, will serve as ABX’s scientific advisor to support the advancement of the technology.
Recently, AI-based de novo antibody design technology has advanced to the point where antibodies can be designed with atomic-level precision using only target antigen information. However, the process of actually validating the vast pool of designed candidates on a large scale has been identified as a bottleneck. As the throughput and cost of standard experimental systems have failed to keep pace with the speed of design, de novo design has been deemed possible but not yet a “mature” technology.
The collaboration between PROTEINA CO., LTD. and ABX focuses on resolving this bottleneck. Under a national R&D project funded by the Ministry of Science and ICT, Professor Baek’s team is developing AbGPT-3D, a multimodal antibody design AI platform. Once this platform designs candidates, PROTEINA CO., LTD. uses its proprietary single-molecule protein interaction analysis platform (SPID) to rapidly and extensively validate antibody-antigen interactions and generate relevant data.
The company explains that it has implemented a so-called “Lab-in-the-Loop” new drug development model—comprising design, high-throughput validation, and retraining—at the world’s highest level. PROTEINA CO., LTD. added that through this approach, it recently became the first in the world to elucidate the core principles of antibody production within mammalian cells.
The newly established subsidiary, ABX, aims to continuously enhance the performance of its AI models to achieve a de novo antibody design success rate that surpasses that of leading global models. PROTEINA CO., LTD., which has secured numerous domestic and international pharmaceutical and biotech partners, will be solely responsible for business development. Revenue generated through de novo design will be shared between the two companies; while roles in AI model development, commercialization, and validation are clearly divided, PROTEINA CO., LTD., as the majority shareholder, will retain overall leadership of the business.
PROTEINA CO., LTD. cited the “Data Flywheel” as the core competitive advantage of this investment. The company explained that the more the AI model is applied to actual projects, the more new antibody-antigen interaction data is accumulated through the SPID platform; this accumulated data, in turn, leads to improved AI model performance, creating a virtuous cycle.
Yoon Tae-young, CEO of PROTEINA CO., LTD., stated, “The final piece of the puzzle in AI-driven antibody drug development is de novo antibody design, and large-scale validation—which has been the biggest bottleneck to date—is the area where PROTEINA CO., LTD. excels.” He added, “Together with ABX, we will complete the closed-loop of design, validation, and retraining to leap forward as a global AI-driven drug development company.”