Phylo 與 Bayer 合作，將 Biomni Lab AI 智能體平台用於藥物研究

Phylo, Inc. announced on October 22, 2026, a partnership with Bayer to deploy Biomni Lab, Phylo’s agentic AI platform, to accelerate scientific research and evaluation, with Bayer scientists gaining full access for literature and omics evidence workflows in drug discovery.

The Bayer Partnership

Under the agreement, Phylo will provide Bayer with full access to its agentic platform. Phylo says the platform helps Bayer scientists bring together scientific literature, biomedical data, and analytical methods in one environment, reducing the manual effort required to assemble and assess evidence. The Bayer workflows will focus on assembling literature and omics evidence for drug discovery.

Sai Jasti, Senior Vice President, Head of Data Science & Artificial Intelligence, Bayer Pharmaceuticals R&D, said the partnership “reflects our belief that AI should expand scientific curiosity, not just speed up routine work.” Jasti said an AI partner that applies rigor at every step allows Bayer scientists to consider more ideas and pose more ambitious questions, with the stated aim of bringing new medicines to patients faster.

Phylo says the platform’s workflows are inspectable: each result preserves its underlying context, so researchers can review the sources, methods, and execution behind it and apply their own scientific judgment.

Kexin Huang, Ph.D., Co-Founder and CEO of Phylo, said, “For agentic AI to matter in research, it has to earn the trust of the scientists doing the work.” Huang called Bayer’s selection of Biomni Lab a strong endorsement of the rigor and AI capabilities Phylo has built into the platform, and described the company’s vision of an AI-native operating system for biopharma, one where each scientist can direct a team of AI agents against complex research questions. With Bayer, Huang said, Phylo aims to show what becomes possible when AI capabilities meet deep scientific expertise.

The Biomni Lab Platform

Phylo calls Biomni Lab the first Integrated Biology Environment, a category the company defines as a single workspace where scientists can search literature, design experiments, analyze biological data, interpret results, and plan next steps. According to Phylo’s Biomni Lab product page, the platform supports research across drug discovery, genomics, bioinformatics, and multi-omics, and is built for bench scientists, computational biologists, bioinformaticians, translational researchers, and life sciences R&D teams. Individual researchers can use it to explore scientific questions and run analyses, while multidisciplinary teams can collaborate across data, literature, and research workflows.

Phylo’s documentation contrasts Biomni Lab with general-purpose LLMs and AI assistants, describing it as purpose-built for biology and life sciences R&D: scientists define questions, design experiments, execute analyses, and interpret results alongside AI agents. For reliability, the platform provides citations and links to original sources when referencing scientific literature and preserves the tools, code, methodology, and execution details behind each analysis so researchers can inspect the approach, verify results, and reproduce the workflow. The documentation cautions that, as with any AI-assisted scientific work, researchers should review results using appropriate scientific judgment.

Phylo’s enterprise tier, Biomni Lab for Enterprise, provides dedicated infrastructure, custom agents, advanced security features, and priority support for larger life sciences organizations, and it can connect with an organization’s proprietary research data, tools, and processes across multidisciplinary workflows. The managed product shares its research lineage with the open-source Biomni project, which originated at Stanford and offers community code that scientists and developers can self-host and customize; Biomni Lab adds a managed research environment, collaboration features, scalable infrastructure, and continuously updated capabilities.

Phylo says Biomni Lab supports drug discovery across early discovery, target validation, lead research, preclinical analysis, and clinical evidence generation; listed uses include synthesizing disease biology, analyzing omics data, prioritizing and evaluating targets, investigating biological structures and candidates, and reviewing clinical evidence. The genomics coverage spans human genetics research, variant annotation, causal gene prioritization, RNA-seq analysis, single-cell analysis, and spatial transcriptomics, with researchers able to define a scientific objective in natural language and then inspect the resulting methodology, code, visualizations, and interpretation.

Bioinformatics and multi-omics workflows cover genomics, transcriptomics, proteomics, single-cell data, and spatial biology; researchers can plan and execute computational analyses, work with datasets, interpret results, create visualizations, and generate reports while preserving the code, sources, and execution details behind the work.

R&D leaders can request an enterprise demonstration of Biomni Lab at contact@phylo.bio.