Elix and LINC Launch World's First Commercial AI Drug Discovery Platform Using Federated Learning from 16 Pharmaceutical Companies
核心洞察
Elix (搜索) and LINC (搜索) have commercialized the world's first AI drug discovery platform incorporating federated learning models trained on data from 16 pharmaceutical companies, addressing the critical data scarcity challenge in AI drug discovery.
The platform leverages over 1 million compounds and 10 million data points through the kMoL (搜索) federated learning library, enabling collaborative AI model development while preserving confidential pharmaceutical data.
The initiative emerged from Japan's AMED DAIIA project and includes AI models for on/off-target prediction, ADMET prediction, and molecular generation, with several pharmaceutical companies already adopting the Elix Discovery (搜索)™ platform.
Elix (搜索), Inc. and the Life Intelligence Consortium (搜索) (LINC (搜索)) have achieved a pharmaceutical industry milestone by launching the world's first commercial AI drug discovery platform that incorporates federated learning models trained on data from 16 pharmaceutical companies. The breakthrough addresses one of the most significant challenges in AI-driven drug discovery: data scarcity.
Revolutionary Federated Learning Approach
The key innovation lies in federated learning technology, which enables multiple pharmaceutical companies to collaboratively develop AI models without disclosing their confidential data externally. Elix (搜索), in partnership with Kyoto University's Department of Biomedical Data Intelligence, developed the federated learning library kMoL (搜索), which serves as the foundation for this unprecedented collaboration.
"Data scarcity remains one of the biggest challenges in AI drug discovery," said Shinya Yuki, Ph.D., Co-Founder and CEO of Elix (搜索). "By jointly developing federated learning technology, kMoL (搜索), with the Department of Biomedical Data Intelligence, Graduate School of Medicine, Kyoto University, we have created a system that enables learning from the data held by 16 pharmaceutical companies while preserving the confidentiality of data such as compound structures."
Unprecedented Scale of Training Data
The federated learning-based AI models were trained on an unprecedented scale of pharmaceutical data. According to Teruki Honma, Ph.D., Team Director at RIKEN (搜索)'s Molecular Design Control Research Team, "DAIIA's predictive AIs were built using chemical structure data provided by pharmaceutical companies, and training on structural data covering more than 1 million compounds and over 10 million data points, is unprecedented on a global scale."
The AI models encompass three critical areas of drug discovery:
- On/off-target prediction
- ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) prediction
- Molecular generation
AMED DAIIA Project Foundation
The development originated from the "Development of a Next-generation Drug Discovery AI through Industry-academia Collaboration" (DAIIA) project, launched in FY2020 under Japan's Agency for Medical Research and Development (AMED). The ambitious project involved 17 pharmaceutical companies, research institutes including RIKEN (搜索), Kyoto University, and Nagoya University, along with approximately 10 IT companies with AI expertise.
The project concluded in March 2025, but Elix (搜索) and LINC (搜索) have ensured continuity by commercializing the innovative models and mechanisms developed during the program. The commercial platform launched in April 2025, marking the transition from research to real-world application.
Industry Collaboration and Adoption
Several pharmaceutical companies have already adopted the Elix Discovery (搜索)™ platform, with expectations for further expansion. Initially, the primary users will be the pharmaceutical companies that participated in DAIIA, but the platform is designed to accommodate additional companies, which will further enhance the AI models' accuracy and usability.
Yasushi Okuno, Ph.D., Representative Director of LINC (搜索) and Professor at Kyoto University, emphasized the significance of this collaborative approach: "Multiple pharmaceutical companies will continue to share data across the industry through federated learning, aiming to develop highly accurate AI. In an industry where the pursuit of individual corporate profit often takes precedence, the effort by each of these companies to share data for the benefit of patients and to build and utilize high-performance drug-discovery AI is profoundly meaningful."
Advanced AI Capabilities
The platform incorporates sophisticated AI technologies, including plans to extend ChemTS (搜索) and incorporate advanced functions such as DyRAMO (搜索), which enables efficient multi-objective optimization. This advancement will allow researchers to create novel compounds and evaluate their activity profiles with higher accuracy and speed than previously possible.
The federated learning system was made possible by a dedicated infrastructure capable of stable federated learning, allowing collaborative model development while preserving confidentiality—a critical requirement in the competitive pharmaceutical industry.
Strategic Positioning and Future Outlook
Elix Discovery (搜索)™ is positioned to become the de-facto standard for AI drug discovery in Japan, with potential for global expansion. The platform's unique approach to data sharing while maintaining confidentiality represents a new paradigm in pharmaceutical collaboration.
"This federated learning based initiative is just the starting point for further progress," noted Yuki. "By encouraging even greater participation and data contributions from pharmaceutical companies, we aim to further expand and strengthen this initiative, enhancing our contribution to the pharmaceutical industry as a whole and ultimately to patients."
The commercialization represents a rare success story in translating government-funded research into practical industry applications, with direct implications for accelerating drug discovery processes and potentially reducing the enormous costs and time traditionally associated with pharmaceutical development.