基于症候-疾病-靶点/通路中医药数据集

针对中医药食同源领域处方难溯源、资源分散、机理验证断裂等痛点，依托超算算力与数据治理能力，搭建一体化研发体系。汇聚整合药典及多类中医、保健食品数据库，构建中药 - 成分 - 靶点知识图谱与疾病靶点通路标准化底库；基于 Qwen 基座微调专属中医大模型，实现辨证方药智能推荐。同时划分配方类型、量化优选评分，结合 RAG 证据回溯与分子对接验证，形成推荐 - 解释 - 验证可追溯闭环，支撑药食同源及保健品配方研发与机理解析。

Addressing the pain points in the field of traditional Chinese medicine (TCM) and food homology, such as difficult traceability of prescriptions, scattered resources, and broken mechanism verification, an integrated R&D system is built relying on supercomputing power and data governance capabilities. This system aggregates and integrates pharmacopoeias, multiple types of TCM and health food databases, and constructs standardized foundational databases including TCM-component-target knowledge graphs and disease target pathway repositories. Based on the Qwen base model, a dedicated TCM large language model (LLM) is fine-tuned to enable intelligent recommendation of syndrome-differentiated prescriptions. Furthermore, it categorizes prescription types, quantifies optimal selection scores, and combines retrieval-augmented generation (RAG)-based evidence backtracking and molecular docking verification to form a traceable closed-loop covering recommendation, explanation and verification, supporting the R&D and mechanism analysis of TCM and food homology as well as health food formula development.

搜集汇总

背景与挑战

背景概述

该数据集针对中医药食同源领域处方难溯源、资源分散等痛点，整合药典及多类数据库，构建了中药-成分-靶点知识图谱与疾病靶点通路标准化底库。它基于微调的中医大模型实现辨证方药智能推荐，并结合RAG证据回溯与分子对接验证，形成可追溯的闭环体系，以支撑配方研发与机理解析。