Preprint

Biomni is a general-purpose biomedical AI agent designed to autonomously execute a wide spectrum of research tasks across diverse biomedical subfields. It employs an action discovery agent to mine tools, databases, and protocols from tens of thousands of publications across 25 biomedical domains, creating the first unified agentic environment (Biomni-E1). Its generalist agentic architecture (Biomni-A1) integrates LLM reasoning with retrieval-augmented planning and code-based execution, dynamically composing complex workflows without predefined templates. Systematic benchmarking demonstrates strong zero-shot generalization across heterogeneous tasks including causal gene prioritization, drug repurposing, rare disease diagnosis, microbiome analysis, and molecular cloning.

Biomedical research underpins progress in our understanding of human health and disease, drug discovery, and clinical care. However, with the growth of complex lab experiments, large datasets, many analytical tools, and expansive literature, biomedical research is increasingly constrained by repetitive and fragmented workflows that slow discovery and limit innovation. Here, we introduce Biomni, a general-purpose biomedical AI agent designed to autonomously execute a wide spectrum of research tasks across diverse biomedical subfields. To systematically map the biomedical action space, Biomni first employs an action discovery agent to create the first unified agentic environment, mining essential tools, databases, and protocols from tens of thousands of publications across 25 biomedical domains. Built on this foundation, Biomni features a generalist agentic architecture that integrates LLM reasoning with retrieval-augmented planning and code-based execution, enabling it to dynamically compose and carry out complex biomedical workflows entirely without relying on predefined templates or rigid task flows. Systematic benchmarking demonstrates that Biomni achieves strong generalization across heterogeneous biomedical tasks including causal gene prioritization, drug repurposing, rare disease diagnosis, microbiome analysis, and molecular cloning without any task-specific prompt tuning. Real-world case studies further showcase Biomni's ability to interpret complex, multi-modal biomedical datasets and autonomously generate experimentally testable protocols.

High-dimensional Mendelian randomization screen of 231 dietary traits and 731 immune phenotypes for insomnia, with cross-release follow-up in FinnGen R12/R13 and population-based validation in NHANES and CHARLS. Prioritized omelette-related intake (protective, OR 0.77) and mixed-fruit (risk, OR 1.29) dietary signals, plus CD33- and HLA-DR-related immune traits. CRP associated with sleep problems in both NHANES and CHARLS. Exploratory cross-design analysis — does not establish causal mechanism.

High-dimensional Mendelian randomization (MR) screens can prioritize candidate dietary and immune pathways for insomnia, but their interpretation is constrained by multiple testing, cross-dataset instability, and limited correspondence between genetic constructs and measured population variables. We conducted an exploratory cross-design analysis that combined MR screening of 231 dietary traits and 731 immune phenotypes, targeted cross-release genetic follow-up in FinnGen R12 and R13, and population-based analyses in NHANES and CHARLS. The targeted R13 follow-up prioritised an omelette-related dietary signal (OR 0.773, 95% CI 0.651-0.917), a mixed-fruit signal (OR 1.285, 95% CI 1.102-1.498), and CD33- and HLA-DR-related immune-cell traits. In NHANES, mapped omelet/scrambled-egg intake was associated with lower odds of sleep problems (OR 0.746, FDR=0.033) and doctor-reported sleep disorder (OR 0.313, FDR=0.008). Higher CRP was associated with sleep problems in NHANES (OR 1.192, FDR=0.001) and restless sleep in CHARLS (OR 1.097, FDR<0.001).