[E-Daily Reporter Yun Junghoon ] SK Group’s information and communications technology (ICT) and biotech affiliates have joined forces to achieve tangible results in the development of AI-powered treatments for intractable cancers. By completing the initial stages of new drug development—which typically take several years—in just five months, they have demonstrated their potential as game-changers in the “AI Bio” era.

SKTelecom(017670)announced on the 15th that it had applied its proprietary AI technology to a joint research project with SK BIOPHARMACEUTICALS and identified two early-stage lead compounds (hits) that can be used in the development of targeted therapies for intractable cancers.

A “hit” is the first step in new drug development and refers to a candidate compound that effectively binds to the target protein responsible for causing the disease. Through this research, the two companies mass-generated and screened candidate binders that bind to “ROR1,” a protein found on the surface of cancer cells, and ultimately confirmed through laboratory validation that two of these candidates possess the efficacy required of a hit.

The key to this achievement lies in the “innovative reduction of time and cost.” They drastically shortened the initial research phase of new drug development—which typically took 1 to 2 years using traditional methods—to just 5 months. This represents a time savings of at least 60% compared to conventional methods.

This ultra-rapid development was made possible by SKT’s advanced machine learning and reinforcement learning (RL) technologies. To overcome the limitations of the early stages of new drug development, where data is scarce, SKT introduced a machine learning technique that combines various “fragments”—small protein pieces—in diverse ways.

In particular, by applying a “reinforcement learning” algorithm that provides high rewards whenever the AI identifies structurally stable combinations, the company was able to rapidly identify the optimal structure for the new binder. Furthermore, SKT fully mobilized its ultra-high-performance GPU (graphics processing unit) infrastructure to process numerous candidate structures in parallel, thereby maximizing analysis speed.

The decision to target the “ROR1” protein—which has drawn significant attention from both academia and industry—is also highly significant. ROR1 is a protein overexpressed on the surfaces of various cancer cells, including those associated with blood cancers and solid tumors, and is considered a key to developing targeted anticancer drugs that selectively kill cancer cells without harming normal cells.

This collaboration is regarded as a prime “One SK” best practice, demonstrating the synergy created by combining the core AI technology of SKT—an ICT affiliate of the SK Group—with SK BIOPHARMACEUTICALS’ unparalleled expertise in new drug development.

Building on this success, the two companies plan to further deepen their collaboration. Their vision is to expand their scope beyond simple compound discovery to encompass the broader field of bio-AI, including the advancement of AI platforms.

Cho Dong-yeon, Head of AI Convergence at SKTelecom, said, “Based on this achievement, we are also exploring ways to expand the scope of our technological collaboration across the entire bio-AI field, including the development of a bio-specialized large language model (LLM) utilizing our proprietary AI foundation model.”