Agentic AI Reveals Conserved Protein Networks Across Human Tissues

Runyu Guan, Dehao Wu, Qiqi Xie, Yang Li, Haohan Wang· September 1, 2026 View original

Key takeaways

  • Agentic AI can uncover complex biological relationships across diverse datasets that are inaccessible to traditional methods.
  • Conserved protein co-abundance clusters across tissues can reveal shared disease mechanisms and therapeutic targets.
  • The framework provides a global, comparable landscape of protein interactions, serving as a valuable hypothesis-generating resource.
  • Non-obvious tissue relationships, like skin-bone marrow, can yield significant biological insights.

Who benefits

PharmaceuticalsBiotechnologyHealthcareMedical Research

Summary

An LLM-agent framework has been developed to identify conserved protein co-abundance clusters across 41 human tissues, uncovering shared disease mechanisms and therapeutic targets previously inaccessible through single-dataset analysis. The framework integrates diverse biological evidence to construct and compare tissue-specific protein networks, highlighting non-obvious relationships and generating mechanistic hypotheses.

Researchers have introduced an innovative LLM-agent framework designed to conduct large-scale, evidence-grounded comparisons of protein co-abundance networks across various human tissues. This system moves beyond traditional, pre-selected tissue pair studies by systematically exploring all possible combinations, integrating data from expression atlases, protein interaction databases, pathway annotations, disease catalogs, and scientific literature. Applying this framework to 820 pairwise combinations of 41 human tissues and fluids, the team identified 1,833 conserved co-abundance clusters across 406 tissue pairs. Key findings include the broad connectivity of tissues like colon, synovial fluid, blood, cerebrospinal fluid, and bone marrow, with bone marrow dominating the most cluster-rich pairs. The analysis also revealed unexpected biological insights, such as the skin-bone marrow pair exhibiting more significant relationships than the anatomically adjacent bone-bone marrow pair. Furthermore, specific clusters, like those in colon-breast involving cancer-relevant processes, and mechanistic hypotheses, such as a brain-gut axis related to serotonin and a liver-bone marrow stress-response axis, were generated. This work provides a comprehensive resource for exploring disease mechanisms and potential therapeutic targets.

Why it matters

This agentic AI approach offers a powerful new method for biological discovery, enabling the identification of complex, systemic disease mechanisms and novel therapeutic targets that span multiple tissues, which is crucial for drug development and personalized medicine.

How to implement this in your domain

  1. 1Explore integrating similar LLM-agent frameworks into bioinformatics pipelines for large-scale data analysis.
  2. 2Collaborate with AI researchers to adapt this methodology for identifying cross-domain patterns in other complex datasets.
  3. 3Utilize the generated landscape of conserved protein co-abundance to prioritize research into specific disease pathways.
  4. 4Investigate the identified non-obvious tissue relationships for potential diagnostic biomarkers or therapeutic interventions.

Original post by Runyu Guan, Dehao Wu, Qiqi Xie, Yang Li, Haohan Wang

"arXiv:2608.28990v1 Announce Type: new Abstract: Protein co-abundance clusters preserved across tissues can reveal shared disease mechanisms and candidate therapeutic targets, particularly when proteins implicated in organ-confined diseases converge in peripheral or accessible tis…"

View on X

Originally posted by Runyu Guan, Dehao Wu, Qiqi Xie, Yang Li, Haohan Wang on X · view source

Want to go deeper?

Turn these trends into skills with Learnijoy's hands-on AI & tech courses.

Explore courses