Yuzheng Gu, Haoyu Wang, Jinlong Yang, Tao Zeng, Hewei Liang, Wenxin He, Mengmeng Wang, Zhinan Wu, Liye Yang, Yusheng Xu, Juan Zhao, Yuning Zhang, Yuliang Dong, Yiyi Zhong, Haifeng Zhang, Jinhong Wang, Xing Rao, Yangfeng Wen, Xiaofan Sun, Karsten Kristiansen, Yu Tian, Xin Tong, Ya Wang, Juan Yang, Fushu Liu, Zejun Yang, Wangsheng Li, Bo Wang, Peng Gao, Jun Xu, Yinglei Miao, Xin Jin, Chuanyu Liu, Xun Xu, Yang Sun, Feng Zhao, Liang Xiao, Yuanqiang Zou
Metabolic interactions govern gut microbiome assembly, yet their functional rules remain obscured by genomic incompleteness and fragmentation. Here, we leverage 1,150 complete genomes to construct genome-scale metabolic models, demonstrating that draft assemblies introduce systematic artifacts and omit critical transport functions. We observe that genomic traits and niche specialization, rather than random association, shape microbial metabolic competition and complementarity. Interaction asymmetry stratifies strains into four ecological groups, including active players, resource predators, resource utilizers, and resource contributors, with distinct signatures of metabolite exchange, competition, and secondary metabolism. In inflammatory bowel disease, these groups show subtype-specific temporal instability, and group-specific dysbiosis predicts clinical phenotypes better than the whole-community profiles. Keystone features derived from integrated metabolic interaction and co-occurrence networks also improve cross-validated disease classification. Together, these findings connect genome completeness with microbial ecological organization and provide a framework for linking metabolic interactions to microbiome-associated disease.