Gelonghui, January 14th: Dongyangguang Pharmaceutical (06887.HK) announced that the group is further deepening its AI strategic layout by launching an AI-driven intelligent R&D platform for PROTAC mechanisms. This platform achieves a complete process from data infrastructure to 'dry-wet closed-loop,' with data scale and structural granularity significantly surpassing existing public databases. It provides a comprehensive and systematic data foundation for AI-driven rational PROTAC design, accelerating the rational design and clinical translation of PROTAC lead compounds.
Traditional small-molecule drugs rely on well-defined binding pockets in target proteins. However, about 80% of human disease-related proteins are considered 'undruggable' due to the lack of suitable structural features. PROTAC (Proteolysis Targeting Chimeras) technology leverages the cell's ubiquitin-proteasome system to induce target protein degradation and has already covered over 200 targets, offering advantages such as catalytic efficiency, overcoming drug resistance, and low toxicity. However, its dependence on ternary complex formation efficiency and spatial orientation makes it complex to design, compounded by fragmented experimental data and a lack of unified standards, limiting the application of AI models and resulting in heavy reliance on empirical trial and error in research and development.
The group’s PROTAC intelligent R&D platform aims to build a high-quality, structured PROTAC-specific database (HEC-PROTACDB), systematically integrating molecular structures, degradation activity, physicochemical properties, and pharmacokinetic parameters from clinical pipelines, patents, and literature. It also conducts detailed modular analysis of POI ligands, Linkers, and E3 ligands. The database currently contains more than 30,000 PROTAC molecules, covering over 500 POI ligands, more than 1,000 types of Linkers, over 200 E3 ligands, and more than 150 targets, with both data scale and structural granularity significantly superior to existing public databases. Based on this, an AI molecule generation and screening platform tailored for PROTAC mechanisms has been developed, forming a 'data-generation-computational validation-experimental feedback' dry-wet closed loop, accelerating the rational design and clinical translation of PROTAC lead compounds.
The group continues to invest heavily in AI R&D platform development, deepening strategic partnerships in AI-driven drug discovery. Six proprietary models have been built, covering the full chain of drug discovery from 'molecular design-activity transition-metabolism simulation-toxicity avoidance-formulation design.' Collaborations with industry leaders like Beijing Deep Potential Technology Co., Ltd. and Huawei Cloud Computing Technology Co., Ltd. have resulted in a comprehensive set of AI scientific computing models for drug discovery, creating a new ecosystem for AI + pharmaceuticals. The company’s first AI-driven small-molecule drug, HEC169584, used for treating metabolic-associated steatohepatitis (MASH), has entered Phase I clinical trials. Preclinical studies show that HEC169584 exhibits high in vitro activity against THR-β cells, strong liver targeting, and reduces impact on the thyroid axis, heart, and other tissues.
