CoChat
Sign Up
Cited by 62Artificial Intelligence in Healthcare and EducationBiomedical Text Mining and OntologiesMachine Learning in Healthcare

An agentic system for rare disease diagnosis with traceable reasoning

Open access
Zhao, Weike, Wu, Chaoyi, Fan, Yanjie, Qiu, Pengcheng, Zhang, Xiaoman, Sun, Yuze, Zhou, Xiao, Zhang, Shuju +7 more·2026·Nature
DOI 10.1038/s41586-025-10097-9PMC PMC12999473
Open this paper in CoChat — free accountOpen PDF

Abstract

Rare diseases affect more than 300 million people worldwide1–3, yet timely and accurate diagnosis remains an urgent challenge1,3–5. Patients often endure a prolonged ‘diagnostic odyssey’ exceeding 5 years, marked by repeated referrals, misdiagnoses and unnecessary interventions, leading to delayed treatment and substantial emotional and economic burden4,5. Here we present DeepRare—a multi-agent system for rare disease differential diagnosis decision support6–8 powered by large language models, integrating more than 40 specialized tools and up-to-date knowledge sources. DeepRare processes heterogeneous clinical inputs, including free-text descriptions, structured human phenotype ontology terms and genetic testing results to generate ranked diagnostic hypotheses with transparent reasoning linked to verifiable medical evidence. Evaluated across nine datasets from literature, case reports and clinical centres across Asia, North America and Europe spanning 14 medical specialties, DeepRare demonstrates exceptional performance on 2,919 diseases. In human-phenotype-ontology-based tasks, it achieves an average Recall@1 of 57.18%, outperforming the next best method by 23.79%; in multi-modal tests, it reaches 69.1% compared with Exomiser’s 55.9% on 168 cases. Expert review achieved 95.4% agreement on its reasoning chains, confirming their validity and traceability. Our work not only advances rare disease diagnosis but also demonstrates how the latest powerful large-language-model-driven agentic systems can reshape current clinical workflows. DeepRare—a multi-agent system for rare disease differential diagnosis decision support powered by large language models, integrating specialized tools and up-to-date knowledge sources—has the potential to reduce healthcare disparities in rare disease diagnosis.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41586-025-10097-9

Cites

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Multi-agent systems distribute clinical reasoning across specialized agents such as history-takers, evidence retrievers, diagnosticians, and decision-makers, with inter-agent communication channels serving as inherently traceable explanation mechanisms.

    “These multi-agent systems [66,77,88,102,118] distribute clinical reasoning across specialized agents such as history-takers, evidence retrievers, diagnosticians, and decision-makers, with inter-agent communication channels serving as inherently traceable explanation mechanisms.”

    From Explainability of decoder-only clinical large language models: A scoping review (2026) · cites this paper as [102]

    Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

    Experimental — accuracy not yet measured

  • A minority of studies engaged clinicians to evaluate explanations, while most relied entirely on automated metrics or proxies.

    “A minority of studies engaged clinicians to evaluate explanations [67,83,87,102,104], while most relied entirely on automated metrics or proxies.”

    From Explainability of decoder-only clinical large language models: A scoping review (2026) · cites this paper as [102]

  • Others expose intermediate reasoning steps such as agentic traces or structured decoding chains that allow clinicians to audit the logic.

    “Others expose intermediate reasoning steps such as agentic traces, structured decoding chains etc that allow clinicians to audit the logic , as in agentic systems [66,77,88,102,118] and structureddecoding approaches [65,101].”

    From Explainability of decoder-only clinical large language models: A scoping review (2026) · cites this paper as [102]

  • Twelve studies (17.4%) engaged healthcare professionals to evaluate generated explanations.

    “Twelve studies (17.4%) engaged healthcare professionals to evaluate generated explanations”

    From Explainability of decoder-only clinical large language models: A scoping review (2026) · cites this paper as [102]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

The citation lists haven’t been fetched yet

References 46 · Cited by 62

Research works printed but not linked: 56

  1. [1]Valdez, R., Ouyang, L. & Bolen, J. Public health and rare diseases: oxymoron no more. Prev. Chronic Dis. 13, E05 (2016). 26766846 PMC4714940 · DOI
  2. [2]Nguengang Wakap, S. et al. Estimating cumulative point prevalence of rare diseases: analysis of the Orphanet database. Eur. J. Human Genet. 28, 165–173 (2020). 31527858 PMC6974615 · DOI · Paper page
  3. [3]The Lancet Global Health. The landscape for rare diseases in 2024. Lancet Glob. Health 12, e341 (2024). 38365397 · DOI
  4. [4]Schieppati, A., Henter, J. I., Daina, E. & Aperia, A. Why rare diseases are an important medical and social issue. Lancet 371, 2039–2041 (2008). 18555915 · DOI
  5. [6]Chen, X. et al. RareAgents: autonomous multi-disciplinary team for rare disease diagnosis and treatment. Preprint at https://arxiv.org/abs/2412.12475v3 (2024). · arXiv
  6. [7]Zou, J. & Topol, E. J. The rise of agentic AI teammates in medicine. Lancet 405, 457 (2025). 39922663 · DOI · Paper page
  7. [9]The Lancet Hope for rare diseases. Lancet 404, 1701 (2024). 39488398 · DOI
  8. [11]Chen, X. et al. Rarebench: can LLMs serve as rare diseases specialists? In Proc. 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining 4850–4861 (2024; 10.1145/3637528.3671576). · DOI · Paper page
  9. [13]Lunke, S. et al. Integrated multi-omics for rapid rare disease diagnosis on a national scale. Nat. Med. 29, 1681–1691 (2023). 37291213 PMC10353936 · DOI
  10. [14]Kernohan, K. D. & Boycott, K. M. The expanding diagnostic toolbox for rare genetic diseases. Nat. Rev. Genet. 25, 401–415 (2024). 38238519 · DOI
  11. [15]The Lancet Digital Health Exploring electronic health records to study rare diseases. Lancet 7, e103 (2025). 39890237
  12. [16]Macken, W. L. et al. Specialist multidisciplinary input maximises rare disease diagnoses from whole genome sequencing. Nat. Commun. 13, 6324 (2022). 36344503 PMC9640711 · DOI
  13. [17]Khalife, L. et al. Multidisciplinary clinical and translational approach for optimizing management for complex and rare conditions using Kabuki syndrome as example. Rare 1, 100008 (2023). · DOI
  14. [18]Dawkins, H. J. S. et al. Progress in rare diseases research 2010–2016: an IRDiRC perspective. Clin. Translat. Sci. 11, 11 (2017). PMC5759730 28796411 · DOI
  15. [19]Qiu, J. et al. LLM-based agentic systems in medicine and healthcare. Nat. Mach. Intell. 6, 1418–1420 (2024). · DOI · Paper page
  16. [20]Lee, Y., Ferber, D., Rood, J. E., Regev, A. & Kather, J. N. How AI agents will change cancer research and oncology. Nat. Cancer 5, 1765–1767 (2024). 39690222 · DOI
  17. [21]Talebirad Y. & Nadiri A. Multi-agent collaboration: harnessing the power of intelligent LLM agents. Preprint at https://arxiv.org/abs/2306.03314v1 (2023). · arXiv
  18. [22]Zheng, Q. et al. How well can modern LLMs act as agent cores in radiology environments? Preprint at https://arxiv.org/abs/2412.09529v3 (2024). · arXiv
  19. [25]Zheng Q. et al. End-to-end agentic RAG system training for traceable diagnostic reasoning. Preprint at https://arxiv.org/abs/2508.15746v1 (2025). · arXiv
  20. [26]Qiu P. et al. Evolving diagnostic agents in a virtual clinical environment. Preprint at https://arxiv.org/abs/2510.24654v1 (2025). · arXiv
  21. [27]DeepSeek-AI, Liu, A. et al. DeepSeek-V3 technical report. Preprint at https://arxiv.org/abs/2412.19437v2 (2024). · arXiv
  22. [28]Mao, X. et al. A phenotype-based AI pipeline outperforms human experts in differentially diagnosing rare diseases using EHRs. NPJ Digit. Med. 8, 68 (2025). 39875532 PMC11775211 · DOI
  23. [29]Fujiwara, T., Shin, J.-M. & Yamaguchi, A. Advances in the development of PubCaseFinder, including the new application programming interface and matching algorithm. Hum. Mutat. 43, 734–742 (2022). 35143083 PMC9305291
  24. [30]Achiam J. et al. GPT-4 technical report. Preprint at https://arxiv.org/abs/2303.08774v6 (2023). · arXiv
  25. [32]Reid, M. et al. Gemini 1.5: unlocking multimodal understanding across millions of tokens of context. CoRR (2024).
  26. [35]Wang, B. et al. Baichuan-M1: pushing the medical capability of large language models. Preprint at https://arxiv.org/abs/2502.12671v2 (2025). · arXiv
  27. [36]Wu, C. et al. Towards evaluating and building versatile large language models for medicine. NPJ Digit. Med. 8, 58 (2025). 39865143 PMC11770143 · DOI · Paper page
  28. [37]Kim, Y. et al. MDAgents: an adaptive collaboration of LLMs for medical decision-making. In Proc. Advances in Neural Information Processing Systems37 (eds Globerson, A. et al.) 79410–79452 (Nomis Foundation, 2024).
  29. [38]Philippakis, A. A. et al. The Matchmaker Exchange: a platform for rare disease gene discovery. Hum. Mutat.36, 915–921 (2015). PMC4610002 26295439
  30. [39]Robinson, P. N. et al. Interpretable clinical genomics with a likelihood ratio paradigm. Am. J. Hum. Genet. 107, 403–417 (2020). 32755546 PMC7477017 · DOI
  31. [40]Firth, H. V. et al. DECIPHER: database of chromosomal imbalance and phenotype in humans using Ensembl resources. Am. J. Hum. Genet. 84, 524–533 (2009). 19344873 PMC2667985 · DOI · Paper page
  32. [41]Töpel, T., Scheible, D., Trefz, F. & Hofestädt, R. Ramedis: a comprehensive information system for variations and corresponding phenotypes of rare metabolic diseases. Hum. Mutat. 31, E1081–E1088 (2010). 19953641 · DOI
  33. [43]Ronicke, S. et al. Can a decision support system accelerate rare disease diagnosis? Evaluating the potential impact of Ada DX in a retrospective study. Orphanet J Rare Dis. 14, 69 (2019). 30898118 PMC6427854 · DOI · Paper page
  34. [45]Miller, N., Lacroix, E.-M. & Backus, J. E. B. MedlinePlus: building and maintaining the National Library of Medicine’s consumer health web service. Bull. Med. Libr. Assoc. 88, 11 (2000). 10658959 PMC35193
  35. [46]Smedley, D. et al. Next-generation diagnostics and disease-gene discovery with the Exomiser. Nat. Protoc. 10, 2004–2015 (2015). 26562621 PMC5467691 · DOI · Paper page
  36. [47]Remy, F., Demuynck, K. & Demeester, T. BioLORD: learning ontological representations from definitions for biomedical concepts and their textual descriptions. In EMNLP Findings 2022. Association for Computational Linguistics (ACL) 1454–1465 (2022).
  37. [48]Macleod, M. R. PubMed: http://www.pubmed.org. J. Neurol. Neurosurg. Psychiatry73, 746 (2002). PMC1757358 12438481
  38. [49]Weinreich, S. S., Mangon, R., Sikkens, J. J. & En Teeuw, M. E. and M. C. Cornel. Orphanet: a European database for rare diseases [Article in Dutch]. Ned. Tijdschr. Geneeskd 152, 518–519 (2008). 18389888
  39. [50]Amberger, J. S., Bocchini, C. A., Schiettecatte, F., Scott, A. F. & Hamosh, A. OMIM.org: Online Mendelian Inheritance in Man (OMIM®), an online catalog of human genes and genetic disorders. Nucleic Acids Res. 43, D789–D798 (2015). 25428349 PMC4383985 · DOI · Paper page
  40. [51]Gargano, M. A. et al. The Human Phenotype Ontology in 2024: phenotypes around the world. Nucleic Acids Res. 52, D1333–D1346 (2023). PMC10767975 37953324 · DOI · Paper page
  41. [52]Qiao Jin, W. et al. MedCPT: Contrastive pre-trained transformers with large-scale PubMed search logs for zero-shot biomedical information retrieval. Bioinformatics 39, btad651 (2023). 37930897 PMC10627406 · DOI · Paper page
  42. [53]Robertson, S. & Zaragoza, H. et al. The probabilistic relevance framework: BM25 and beyond. Found Trends Information Retrieval 3, 333–389 (2009). · DOI · Paper page
  43. [55]Siva, N. 1000 Genomes Project. Nature 26, 256 (2008). 18327223
  44. [56]Daniel Taliun, D. N. et al. Sequencing of 53,831 diverse genomes from the NHLBI TOPMed program. Nature 590, 290–299 (2021). 33568819 PMC7875770 · DOI · Paper page
  45. [57]UK10K Consortium. The UK10K project identifies rare variants in health and disease. Nature 526, 82–90 (2015). 26367797 PMC4773891 · DOI · Paper page
  46. [59]Ivan, A. et al. A method and server for predicting damaging missense mutations. Nat. Methods 7, 248–249 (2010). 20354512 PMC2855889 · DOI · Paper page
  47. [60]Ng, P. C. & Henikoff, S. SIFT: predicting amino acid changes that affect protein function. Nucleic Acids Res. 31, 3812–3814 (2003). 12824425 PMC168916 · DOI
  48. [61]Schwarz, J. M., Rödelsperger, C., Schuelke, M. & Seelow, D. MutationTaster evaluates disease-causing potential of sequence alterations. Nat. Methods 7, 575–576 (2010). 20676075 · DOI · Paper page
  49. [62]Sebastian Köhler, M. et al. The Human Phenotype Ontology in 2021. Nucleic Acids Res. 49, D1207–D1217 (2021). 33264411 PMC7778952 · DOI
  50. [63]Yates, T. M., Ansari, M., Thompson, L. & Hunt, S. E. Curating genomic disease-gene relationships with Gene2Phenotype (G2P). Genome Med. 16, 127 (2024). 39506859 PMC11539801 · DOI
  51. [64]Zhao, Z., Jin, Q., Chen, F., Peng, T. & Yu, S. A large-scale dataset of patient summaries for retrieval-based clinical decision support systems. Sci. Data 10, 909 (2023). 38110415 PMC10728216 · DOI · Paper page
  52. [66]Landrum, M. J. et al. ClinVar: public archive of interpretations of clinically relevant variants. Nucleic Acids Res. 44, D862–D868 (2016). 26582918 PMC4702865 · DOI · Paper page
  53. [67]Karczewski, K. J. et al. The ExAC browser: displaying reference data information from over 60,000 exomes. Nucleic Acids Res. 45, D840–D845 (2017). 27899611 PMC5210650 · DOI
  54. [68]Fan, Y. et al. Improving variant prioritization in exome analysis by entropy-weighted ensemble of multiple tools. Clin. Genet. 103, 190–199 (2023). 36309956 · DOI
  55. [69]Kasai, J., Kasai, Y., Sakaguchi, K., Yamada, Y. & Radev, D. Evaluating GPT-4 and ChatGPT on Japanese medical licensing examinations. Preprint at https://arxiv.org/abs/2303.18027v2 (2023). · arXiv
  56. [70]Gemini Team Google, Anil, R. et al. Gemini: a family of highly capable multimodal models. Preprint at https://arxiv.org/abs/2312.11805v5 (2023). · arXiv

Authors
Zhao, W., Wu, C., Fan, Y., Qiu, P., Zhang, X., Sun, Y., Zhou, X., Zhang, S., Peng, Y., Wang, Y., Sun, X., Zhang, Y., Yu, Y., Sun, K., & Xie, W.
Affiliations
  • Shanghai Jiao Tong University (CN)
  • Harvard University (US)
  • Shanghai Artificial Intelligence Laboratory (CN)
Corresponding author
  • Zhang, Ya
  • Yu, Yongguo
  • Sun, Kun
  • Xie, Weidi
ORCID
Venue
Nature
Publisher
Springer Science and Business Media LLC
Published
2026-02-18
Volume
651
Issue
8106
Pages
775-784
ISSN
  • 0028-0836 (ISSN-L)
  • 1476-4687
DOI
10.1038/s41586-025-10097-9
PMC
PMC12999473
OpenAlex
W7130436101
Cited by
62
References
46
Citations by year
2026: 61 · 2025: 1
FWCI
119.96
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature (publishedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University)
Funders
  • Shanghai Municipal Health Commission (2020CXJQ01)
  • National Natural Science Foundation of China (82130015)
  • Shanghai Municipal Commission of Economy and Informatization
Topics
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Biomedical Text Mining and Ontologies
  • Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning in Healthcare
Keywords
rare disease diagnosis, differential diagnosis, multi-agent systems, large language models, agentic systems, diagnostic decision support, traceable reasoning, Human Phenotype Ontology, genetic testing, diagnostic odyssey
MeSH terms
Diagnosis, Differential, Humans, Phenotype, Rare Diseases

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review