KYORIN Pharmaceutical Adopts AI Drug Discovery Platform From Elix
KYORIN Pharmaceutical has adopted Elix Discovery, an AI-based drug discovery platform developed by Tokyo-based Elix. The system is designed to accelerate early-stage research by applying machine learning to compound profiling, molecular generation, and predictive modeling.
Elix Discovery, launched in 2022, provides medicinal chemists with a graphical user interface (GUI) that enables the automated construction of predictive models without requiring coding expertise. Its core technology is built on proprietary molecular structure generation models trained to propose novel scaffolds beyond standard chemical libraries, with the aim of suggesting compounds that chemists might not otherwise conceive. The platform incorporates predictive models that estimate molecular activity, ADMET properties, and physicochemical characteristics, ranging from conventional machine learning to deep learning approaches, with automated optimization to identify the best-performing model for a given dataset.
The Platform
Generative models within Elix Discovery support ligand-based and structure-based drug design, handling use cases such as de novo compound generation, lead optimization, and linker design. The system combines these AI models with physics-based docking simulations accelerated by GPUs, enabling structural optimization even when biological activity data is limited. Docking and pharmacophore modeling can also be integrated with predictive models to refine candidate molecules further.
Datasets used to train the platform’s models include contributions from 16 pharmaceutical companies, covering a broad spectrum of chemical and biological assay data. Users may also upload proprietary or public data to train customized models for specific research programs.
Beyond software, Elix provides consulting and direct support from its researchers, positioning this as a critical factor for successful adoption within medicinal chemistry teams. The company emphasizes openness in sharing methods and knowledge, which it contrasts with the more closed approaches of other AI developers. Elix Discovery is available in both Japanese and English and can be deployed on-premise within client infrastructure or accessed through cloud-based services.
The platform is also being extended through collaborations such as Elix’s recent partnership with PRISM BioLab. That effort focuses on hard-to-drug protein-protein interaction targets, combining Elix’s AI-driven compound design with PRISM’s small molecule scaffolds that mimic structural motifs found in intracellular protein interfaces.
KYORIN stated that the adoption of Elix Discovery is intended to strengthen its internal drug discovery pipeline by integrating AI-driven design with its existing expertise. The company expects the system to accelerate identification of candidate molecules and improve the overall efficiency and quality of research in the development of new medicines.
Topics: AI & Digital
