AI Accelerates Cancer Drug Discovery: SK Telecom and SK Biopharmaceuticals Identify ROR1-Targeting Compounds in Five Months
Key Insights
SK Telecom (search) and SK Biopharmaceuticals (search) used AI to identify two early hit compounds targeting ROR1, a protein implicated in blood cancers (search) and solid tumors (search), in approximately five months.
The AI-driven approach reduced the typical early drug discovery timeline of one to two years by more than 60%, leveraging machine learning and reinforcement learning techniques.
The collaboration combined SK Biopharmaceuticals (search)' drug development expertise with SK Telecom (search)'s AI and GPU computing resources to overcome limitations posed by scarce training data.
SK Telecom (search) and SK Biopharmaceuticals (search) have harnessed artificial intelligence to identify two promising early-stage compounds for a targeted cancer therapy in approximately five months, a timeline that slashes the conventional drug discovery process by more than 60%.
The joint research successfully pinpointed two early hit compounds capable of binding to ROR1, a protein expressed on the surface of cancer cells that has emerged as a compelling therapeutic target for a range of blood cancers (search) and solid tumors (search). SK Telecom (search) announced the results on Wednesday, underscoring the potential of AI to reshape the earliest and often most time-consuming phase of drug development.
Overcoming Data Scarcity with Machine Learning
Early drug discovery typically requires one to two years of iterative screening and optimization. The challenge is compounded when using AI, as conventional approaches often falter due to insufficient training data at this nascent stage.
To address this, SK Telecom (search) deployed machine learning techniques that combine protein fragments in diverse configurations, coupled with reinforcement learning algorithms designed to reward structurally stable molecular designs. The company also leveraged its graphics processing unit (GPU) computing infrastructure to process large numbers of binder candidates simultaneously, enabling rapid in silico prediction of how each candidate might interact with ROR1.
"Binders are molecules designed to attach to specific biological targets, such as proteins found on cancer cells," the companies noted, emphasizing that finding new binders typically requires evaluating multiple factors, including binding strength and molecular stability.
A Collaborative Framework
The partnership drew on complementary strengths: SK Biopharmaceuticals (search) contributed its drug development expertise to devise the overall discovery strategy, while SK Telecom (search) applied its AI capabilities to generate candidate molecules and assess their likelihood of binding to ROR1. The AI models rapidly narrowed the field of candidates before laboratory validation, where two compounds were confirmed as early-stage hits.
The companies stated that the five-month timeline represents a reduction of more than 60% compared to traditional methods, which can stretch to two years for comparable early-stage discovery work.
Implications for Targeted Cancer Therapy
ROR1 has attracted growing interest as a drug target due to its expression on malignant cells and limited presence in normal adult tissues, making it an attractive candidate for therapies aimed at both hematologic malignancies and solid tumors (search). The identification of binders against this target marks an incremental but meaningful step toward developing novel cancer treatments.
The research remains in the early hit identification stage, and the compounds will require further optimization and preclinical validation before any clinical development can begin. Nonetheless, the accelerated timeline demonstrates how AI-driven methodologies may compress the drug discovery cycle, potentially bringing new cancer therapies to patients faster.