{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"Research Square"}],"indexed":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T09:23:29Z","timestamp":1787736209601,"version":"build-2784847793"},"posted":{"date-parts":[[2026,7,14]]},"group-title":"In Review","reference-count":0,"publisher":"Springer Science and Business Media LLC","license":[{"start":{"date-parts":[[2026,7,14]],"date-time":"2026-07-14T00:00:00Z","timestamp":1783987200000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2026,6,16]]},"abstract":"<title>Abstract<\/title>\n                <p>Phenotype discovery\u2014the data-driven identification of clinically or biologically meaningful subgroups\u2014is fundamental to precision medicine, but conventional clustering methods require the number of clusters K to be specified a priori and struggle with heterogeneous, multimodal, or longitudinal data. We introduce the GRACE Cycle (Generate hypothesis, Retrieve evidence, Align, Converge, Evaluate), a general large-language-model (LLM)-assisted framework for phenotype discovery in which a hypothesis, an LLM, and an evidence base are iteratively refined until they agree. The framework discovers K as an output through Graph-of-Thought (GoT) refinement, in which an LLM reads per-cluster summary cards plus a between-cluster similarity matrix and proposes one of three moves\u2014SPLIT, MERGE, or COMMIT\u2014over a spectral-clustering seed. Two technical contributions enable scale: (i) a four-component prompt template integrating pairwise comparison, fairness pre-processing, and structured JSON output, and (ii) a data-feeding strategy that compresses cohorts of 10 4 \u201310 6 entities into context-budget-respecting batches via k-nearest-neighbour graph sampling. We validate GRACE across three heterogeneous phenotyping problems: (1) longitudinal Long COVID subphenotyping in the NIH RECOVER cohort (n = 13,511), where GRACE recovers three clinically distinct subphenotypes (Protected, Responder, Refractory) with bootstrap Jaccard stability &gt; 0.97 that are explained by a single autonomic\/post-viral-fatigue axis (a 25-fold dysautono-mia gradient, dysautonomia adjusted OR = 13.4) and an accompanying collapse of wearable-measured physical activity; (2) motor subphenotyping of Parkinson\u2019s disease from foot-sensor gait wearables (PhysioNet gaitpdb, n = 93), where GRACE discovers two gait subtypes without specifying K that are externally validated against withheld Timed-Up-and-Go (p = 0.002), Hoehn\u2013Yahr stage (p = 0.03), and age; and (3) additional open wearable chronic-disease cohorts processed with the identical pipeline. Across domains, GRACE converges without prior knowledge of K, demonstrating that LLM-guided iterative reasoning offers a domain-agnostic alternative to conventional clustering when the number of phenotypes is unknown.<\/p>","DOI":"10.21203\/rs.3.rs-10048424\/v1","type":"posted-content","created":{"date-parts":[[2026,7,14]],"date-time":"2026-07-14T14:33:29Z","timestamp":1784039609000},"source":"Crossref","is-referenced-by-count":0,"title":["The GRACE Cycle: A General Large-Language-Model Framework for Phenotype Discovery with Unknown Cluster Number"],"prefix":"10.21203","author":[{"given":"Jing","family":"Wang","sequence":"first","affiliation":[{"name":"United States National Library of Medicine"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zorina","family":"Galis","sequence":"additional","affiliation":[{"name":"National Heart Lung and Blood Institute"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tong","family":"Zhang","sequence":"additional","affiliation":[{"name":"University of Illinois Urbana-Champaign"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yiming","family":"Luo","sequence":"additional","affiliation":[{"name":"Columbia University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Amar","family":"Sra","sequence":"additional","affiliation":[{"name":"George Washington University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xing","family":"Niu","sequence":"additional","affiliation":[{"name":"Amazon (United States)"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jie","family":"Shen","sequence":"additional","affiliation":[{"name":"Stevens Institute of Technology"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Qiaomin","family":"Xie","sequence":"additional","affiliation":[{"name":"University of Wisconsin\u2013Madison"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jeremy","family":"Weiss","sequence":"additional","affiliation":[{"name":"United States National Library of Medicine"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","container-title":[],"original-title":[],"link":[{"URL":"https:\/\/www.researchsquare.com\/article\/rs-10048424\/v1","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.researchsquare.com\/article\/rs-10048424\/v1.html","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T09:01:18Z","timestamp":1787734878000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.researchsquare.com\/article\/rs-10048424\/v1"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,14]]},"references-count":0,"URL":"https:\/\/doi.org\/10.21203\/rs.3.rs-10048424\/v1","relation":{},"subject":[],"published":{"date-parts":[[2026,7,14]]},"subtype":"preprint"}}