28

Jilong Liu, Jianwen Zhou, Zhaoqiang Li, Qingyu Xiao, Kaixin Yang, Ziru Chen, Lu Yao, Hengshen Liu, Yibo Miao, Junyu Luo, Yiping Chen, Ruifeng jing, Yixue Li

Guangzhou Laboratory, Guangzhou 510005, Guangdong Province, China

While cloud computing has revolutionized data processing and collaboration in life sciences, significant challenges persist in developing tailored cloud platforms for biomedical research and ensuring seamless interoperability. The Global Alliance for Genomics and Health (GA4GH) Cloud application programming interfaces (APIs) provide a critical framework to address these issues. To facilitate large-scale, cross-institutional biomedical data utilization in China while enhancing global connectivity, we developed Bio-medical big data operating system (Bio-OS) based on GA4GH Cloud standards. Bio-OS is an open-source, cloud-native framework engineered to accelerate the development of interoperable biomedical cloud platforms. We established a centralized registry service to interconnect Bio-OS instances, forming a collaborative resource network. Bio-OS implements GA4GH standards including Workflow Execution Service (WES), Task Execution Service (TES), Data Repository Service (DRS), Tool Registry Service (TRS), and Authentication and Authorization Infrastructure (AAI) Passport, while introducing novel features: Interactive Execution Service (IES) for tertiary analysis support, workspace-centric data management, and AI-driven enhancements. The workspace paradigm organizes research projects encompassing introduction, datasets, workflows, interactive analysis instances and logs. We developed Workspace RO-Crate as a standardized format for workspace export and distribution, complemented by Digger (https://network.miracle.ac.cn/digger) repository for artifact sharing. Enhanced machine and Large Language Model operability through Workspace RO-Crate enabled development of Bio-OS Copilot, transforming Bio-OS into an operable environment for AI agents and enabling the AI4LifeScience paradigm. This leverages a VS Code extension for natural language-based workspace manipulation. These innovations promote adherence to FAIR (Findable, Accessible, Interoperable, Reusable) principles, enhancing reproducibility and collaboration. Multi-institutional deployments validate Bio-OS's efficacy in supporting data-intensive biomedical applications and establishing robust cloud platforms for life sciences research.