Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “LLM”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 19 recordsLinked to original sources
PubMed2026

Dataset Readiness Assessment With Large Language Model (DRAFT-LLM): A Multi-Axis Audit Guided by LLM.

This article details the Dataset Readiness Assessment for Training (DRAFT), a systematic method for determining whether a high-dimensional biological dataset is suitable for developing reliable, equitable (i.e., the extent to which model performance, error patterns, and potential benefits or harms are evaluated and found to be acceptably distributed across relevant demographic, biological, clinical, and contextual subgroups), and scientifically meaningful machine-learning models, and DRAFT Large Language Model (DRAFT-LLM), its optional human-in-the-loop extension for calibrating study-specific audits through structured, critically reviewed LLM guidance. Standard model validation often fails to detect when apparent performance is driven by spurious correlations, technical artifacts, or hidden stratification, leading to irreproducible and inequitable findings. DRAFT-LLM addresses this gap by shifting the focus from model tuning to structured dataset auditing, organized around Support Protocols 1 to 4 that capture the scientific intent, data structure, and governance constraints of a given study. These Support Protocols: (1) elicit and formalize investigator input into a study intake and dataset card; (2) compute standardized dataset statistics and structural summaries suitable for downstream analysis and LLM context; (3) configure the language model using form-based responses, safety guardrails, and governance rules; and (4) generate personalized instructions, prompts, and code templates for running DRAFT audits. Basic Protocols 1 to 3 are instantiated from this support layer for generalization, equity, and stability: they are reusable execution patterns whose concrete behavior is determined by the cards, statistics, and configurations defined in the Support Protocols. DRAFT-LLM and DRAFT are demonstrated in this article through an end-to-end case study on The Cancer Genome Atlas (TCGA). © 2026 Wiley Periodicals LLC. Support Protocol 1: Study intake and dataset card construction Support Protocol 2: Dataset structure and advanced summary statistics for LLM context Support Protocol 3: LLM configuration using structured form responses Support Protocol 4: Generation of personalized instructions for DRAFT audits Basic Protocol 1: Generalization audit Basic Protocol 2: Equity audit Basic Protocol 3: Stability audit.

Large Language Models↗
PubMed2026

scGPA: an LLM-assisted workflow for directional virtual gene perturbation analysis from single-cell transcriptomes.

BACKGROUND: Existing virtual perturbation methods can often infer directional changes by comparing predicted post-perturbation expression profiles with control cells. However, workflows that directly return direction-specific downstream candidate genes together with confidence scores, evidence support and interpretable summaries remain limited. We developed scGPA, an LLM-assisted workflow system for directional single-cell virtual gene perturbation analysis. METHODS: scGPA starts from raw single-cell RNA sequencing data and performs quality control, normalization, dimensionality reduction, clustering and cell-group selection. It then constructs cell-group-specific wild-type regulatory networks using repeated subsampling, principal component regression (PCR)/Ridge-based network inference and CP tensor denoising. Based on these networks, scGPA simulates dose-aware virtual knockdown of the target gene and applies signed perturbation propagation to estimate the magnitude and direction of downstream transcriptional responses. LLM assistance is used for marker-based cell-type annotation, evidence-guided candidate prioritization and user-facing biological summarization. RESULTS: We benchmarked scGPA across five public Perturb-seq datasets and compared its performance with GEARS, scGPT and a random baseline. The overall correct prediction rate of scGPA was 23.0%, exceeding those of GEARS (20.7%), scGPT (15.1%) and the random baseline (13.6%). These results indicate that scGPA achieved a higher correct prediction rate than the two comparator models and the random baseline. We subsequently evaluated scGPA using a public osteosarcoma single-cell dataset and performed qRT-PCR validation in 143B osteosarcoma cells. Among genes with significant experimental changes, scGPA achieved a directional concordance of 76.9%. When all tested downstream genes were counted, 37.0% were directionally correct, 51.9% showed no significant change and 11.1% changed in the opposite direction. CONCLUSIONS: scGPA provides a practical workflow system for predicting and prioritizing direction-specific downstream transcriptional responses after target-gene perturbation. By integrating single-cell regulatory network inference, signed virtual perturbation and LLM-assisted interpretation, scGPA supports target-gene function inference and downstream mechanistic investigation from single-cell transcriptomic data.

Single-Cell Gene Expression Analysis↗
PubMed2026

Integrative evidence-knowledge marker selection enhances LLM-based cell type annotation in single-cell RNA-seq analysis.

BACKGROUND: Cell type annotation is essential for gaining biological insight from single-cell RNA sequencing data, yet manual labeling remains time-consuming and difficult to reproduce. Various computational approaches have been developed to automate this process, and recent studies suggest that large language models can infer cell types with promising accuracy in single-cell analysis. However, most workflows still rely on cluster-specific markers derived from gene expression alone or manual curation. As a result, marker selection can be sensitive to statistical criteria and dataset-dependent bias, which may lead to the selection of less informative genes or missing important markers, while providing limited biological context. RESULTS: To address this limitation, we introduce CELLIA, an LLM-based workflow for automated and robust cell type annotation. CELLIA employs an integrative evidence-knowledge marker selection strategy that combines statistical differential expression criteria with curated tissue-specific marker resources to identify informative marker genes. In benchmarking analyses of 102 cell types, this approach improved agreement with manual annotations. In addition, CELLIA achieved higher agreement in subtype-level analyses of closely related immune populations and was further evaluated in a non-immune stromal subtype setting, covering 25 cell types in total. CONCLUSION: By integrating evidence-knowledge from gene expression with curated biological prior knowledge, CELLIA provides a more stable marker selection and improves the reliability of LLM-cell type annotation.

Cell type annotation↗
PubMed2025

AutoPM3: enhancing variant interpretation via LLM-driven PM3 evidence extraction from scientific literature.

MOTIVATION: Rare diseases affect over 300 million people worldwide and are often caused by genetic variants. While variant detection has become cost-effective, interpreting these variants-particularly collecting literature-based evidence like ACMG/AMP PM3-remains complex and time-consuming. RESULTS: We present AutoPM3, a method that automates PM3 evidence extraction from literatures using open-source large language models (LLMs). AutoPM3 combines a Text2SQL-based variant extractor and a retrieval-augmented generation (RAG) module, enhanced by a variant-specific retriever and fine-tuned LLM, to separately process tables and text. We curated PM3-Bench, a dataset of 1027 variant-publication evidence pairs from ClinGen. On openly accessible pairs, AutoPM3 achieved 86.1% accuracy for variant hits and 72.5% recall for in trans variants-outperforming other methods, including those using larger models. We uncovered the effectiveness of AutoPM3's key modules, especially for variant-specific retriever and Text2SQL, through the sequential ablation study. AutoPM3 located evidence in 76 s, demonstrating that open-source LLMs can offer an efficient, cost-effective solution for rare disease diagnosis. AVAILABILITY AND IMPLEMENTATION: AutoPM3 is implemented and freely available under the MIT license at https://github.com/HKU-BAL/AutoPM3.

Genetic Variation↗
PubMed2025

Pre-Meta: priors-augmented retrieval for LLM-based metadata generation.

MOTIVATION: While high-throughput sequencing technologies have dramatically accelerated genomic data generation, the manual processes required for dataset annotation and metadata creation impede the efficient discovery and publication of these resources across disparate public repositories. Large language models (LLMs) have the potential to streamline dataset profiling and discovery. However, their current limitations in generalizing across specialized knowledge domains, particularly in fields such as biomedical genomics, prevent them from fully realizing this potential. This article presents Pre-Meta, an LLM-agnostic and domain-independent data annotation pipeline with an enriched retrieval procedure that leverages related priors-such as pre-generated metadata tags and ontologies-as auxiliary information to improve the accuracy of automated metadata generation. RESULTS: Validated using five selected metadata fields sampled across 1500 papers, the Pre-Meta assisted annotation experiment-without finetuning and prompt optimization-demonstrates a systemic improvement in the annotation task: shown through a 23%, 72%, and 75% accuracy gain from conventional RAG adoptions of GPT-4o mini, Llama 8B, and Mistral 7B respectively. AVAILABILITY AND IMPLEMENTATION: The code, data access, and scripts are available at: https://github.com/SINTEF-SE/LLMDap.

Metadata↗
PubMed2026

RESCUE: An end-to-end multi-agent LLM system for proactive rare-disease patient screening in the EHR.

BACKGROUND: Rare diseases affect a significant portion of the global population, yet patients often endure a lengthy diagnostic odyssey, frequently missing the opportunity for timely diagnoses with exome or genome sequencing (ES/GS). Existing informatics tools often rely on pre-identified patients or rigid, institution-specific rule sets, failing to address the broader operational question of clinical utility and feasibility. METHODS: We introduce RESCUE (Rare Disease Detection and Escalation Support via a Learning Health System), an end-to-end, multi-agent LLM-powered workflow designed for proactive rare-disease diagnosis across the entire electronic health record (EHR). RESCUE utilizes a team of specialized agents including Ontology, Modeling, Screening, and Review, to automate the screening process to identify candidates for diagnostic testing based on their clinical features. The Ontology Agent classifies clinical data into a four-tier genetic-evidence taxonomy; the Modeling Agent builds a positive-unlabeled (PU) XGBoost classifier to identify potential cases; the Screening Agent applies these models across the EHR population; and the Review Agent evaluates candidates by sampling clinical notes to ensure medical necessity and operational feasibility for genomic testing. RESULTS: Using electronic medical record data from a pediatric hospital, our retrospective evaluation on a holdout set (n=12,591) demonstrates strong discrimination between patients who received diagnostic genomic testing and those who did not (AUC 0.808). Of nearly 500,000 patients in the institutional base, 175,842 met inclusion criteria for screening; among these, RESCUE-flagged candidates were 7.4-fold more likely to receive subsequent genomic assessments compared to controls. Blinded manual chart reviews confirmed that RESCUE identifies previously missed, medically appropriate patients for ES/GS with 80% precision, while simultaneously accounting for prior testing history. CONCLUSIONS: By decoupling expert roles into modular agents, RESCUE offers a flexible, scalable, and adaptable framework for screening patients for rare-disease diagnostic genomic testing. This approach overcomes the limitations of traditional rule-based methods and provides a reproducible, agentic pathway to reduce diagnostic delays and improve patient care at an institutional scale.

Journal Article↗
PubMed2026

Pedagogical Efficacy of LLM-Generated Synthetic Data Versus Real-World Clinical Records: A Randomized Controlled Non-Inferiority Trial.

BACKGROUND: Expert-reviewed clinical cases generated by large language models (LLMs) may supplement case resources in medical education, but their short-term educational performance relative to real-case-derived teaching materials remains uncertain. We compared immediate post-training test performance after teaching with the two types of case materials and assessed non-inferiority against a prespecified margin. METHODS: We conducted a prospective, parallel-group, randomized non-inferiority trial. Through the Wenjuanxing online platform, participants were randomized 1:1 to learn with either real-case-derived teaching cases compiled by clinicians and reviewed by experts or AI-generated clinical cases produced by Gemini 3.0 Pro from fully de-identified matched real cases and reviewed by three senior general surgery specialists with full-professor rank. The primary outcome was the total score on an independent 10-item immediate post-training test (0-10 points), with a prespecified non-inferiority margin of -0.5 points. Secondary outcomes included the training-phase performance score, learning efficiency index, single-item mental effort rating, case realism, and case-source judgment. RESULTS: A total of 403 participants were randomized, of whom 386 were included in the modified intention-to-treat analysis: 192 in the real-case group and 194 in the AI-generated case group. The mean post-training test score was 4.95 (SD, 3.35) in the real-case group and 4.61 (SD, 3.35) in the AI-generated case group. The mean difference (AI-generated minus real-case group) was -0.335 points (95% CI, -1.006 to 0.337). Because the lower bound of the confidence interval was below the prespecified non-inferiority margin of -0.5 points, non-inferiority was not demonstrated (one-sided P = 0.314). No significant between-group differences were observed in the training-phase performance score, learning efficiency index, or single-item mental effort rating. AI-generated cases received lower realism ratings for Level 3 cases. The proportion of participants with at least one high-confidence completely incorrect response was 1.6% in the real-case group and 2.1% in the AI-generated case group. CONCLUSIONS: In this short-term, text-based online case-learning setting, no statistically significant between-group difference was observed in immediate post-training test performance; however, non-inferiority of AI-generated clinical cases relative to real-case-derived teaching materials was not demonstrated.

Humans↗
PubMed2026

A bimodal large language model reduces misalignment in patient education: A double-blinded randomized trial.

BACKGROUND: Effective patient education requires accurate communication aligned with patients' emotional and semantical needs. Text-based large language models (LLMs) lack access to non-verbal cues, which may contribute to misaligned responses. METHODS: We evaluated emotional and semantic misalignment in a text-based LLM using 64,200 utterances from 16,583 patient education cases across six departments and three centers. Dolphin was developed integrating text and audio cues and evaluated through emotion recognition, semantic consistency assessment, branch-level ablations, and a double-blinded randomized trial against a matched text-based LLM comparator (Chinese Clinical Trial Registry: (ChiCTR2500095933). FINDINGS: The text-based LLM showed emotional misalignment in 36.7% of responses and semantic misalignment in 28.3% of cases, with higher misalignment under greater burden. Dolphin outperformed the text-based LLM in emotion recognition accuracy (0.886 vs. 0.713) and semantic consistency (84.9% vs. 82.1%; both adjusted p < 0.001). Ablations supported contribution of audio branches. Dolphin received higher expert ratings than the text-based LLM and human educators (all p < 0.001). In 555 patients, Dolphin was associated with greater patient satisfaction (98.6% vs. 93.8%), suggestion acceptance (76.1% vs. 58.9%; p < 0.001), proactive disclosure (44.6% vs. 26.5%; p < 0.001), and fewer 7-day unplanned recontact (12.9% vs. 22.9%; p = 0.002). No unsafe recommendations or safety events were identified. CONCLUSIONS: Compared with text-based LLM, Dolphin improved emotional-semantic alignment and patient-education outcomes, supporting bimodal alignment as a strategy for reducing misalignment-driven communication failures. FUNDING: National Natural Science Foundation of China, State Key Laboratory Special Fund, and Chinese Academy of Medical Sciences Innovation Fund.

Humans↗
PubMed1978

Clinical evaluation of a sensory feedback device: the limb load monitor.

Based on records of 81 patients who used the LLM, and on questionnaire answers and comments from clinicians, the following can be concluded: 1. The LLM can be operated easily after a minimum of training. It does not break down with extended clinical use when handled properly. 2. The LLM manual provides sufficient information for proper operation and clinical use of the device. 3. The number of patients in a clinic who can benefit from LLM training can be predicted, if consideration is given to the type of facility and the size of the patient population and physical therapy staff. 4. The largest diagnostic group of patients who can benefit from LLM therapy are lower-limb amputees, followed by hemiplegic and orthopedic patients. 5. The general selection criteria outlined initially proved sufficient. A patient who is selected properly can be expected to respond to the feedback signal (i.e., make a weight-bearing adjustment) within the first or second session.

Adolescent↗
PubMed1992

Time course of lipid-laden pulmonary macrophages with acute and recurrent milk aspiration in rabbits.

High levels of lipid-laden macrophages (LLM) in bronchial washings have been associated with food aspiration. We studied the time course of appearance and clearance of LLM in rabbits undergoing either a single milk instillation, five weekly milk instillations or saline (control) instillations into the airways. Cells were obtained by bronchoalveolar lavage of intubated rabbits at uniform time intervals following the single or the last of five milk/saline instillations. LLM semi-quantitative indexes were derived using oil-red-O staining. Significantly elevated indexes were found in both milk groups 6 hr after milk instillation. In the single saline and milk instillation groups the indexes were not different beginning on the 4th day, and indexes from 8 of 9 rabbits had returned to baseline by the 6th day. However, indexes remained significantly elevated up to 17 days in the group receiving weekly milk instillations. Indexes from all rabbits in the repeat milk instillation group remained elevated for 12 days or longer. This group also developed increased numbers of binucleated macrophages. Quantitation of LLM in this model appears to be a sensitive indicator of recurrent lipid aspiration, these cells remaining in the airways for several days after the last aspiration event.

Animals↗
PubMed2025

Integrating expert knowledge into large language models improves performance for psychiatric reasoning and diagnosis.

BACKGROUND AND METHODS: The authors sought to evaluate the performance of common large language models (LLMs) in psychiatric diagnosis, and the impact of integrating expert-derived reasoning on their performance. Clinical case vignettes and associated diagnoses were retrieved from the DSM-5-TR Clinical Cases book. Diagnostic decision trees were retrieved from the DSM-5-TR Handbook of Differential Diagnosis and refined for LLM use. Three LLMs were prompted to provide diagnosis candidates for the vignettes either by directly prompting or using the decision trees. These candidates and diagnostic categories were compared against the correct diagnoses. The positive predictive value (PPV), sensitivity, and F1 statistic were used to measure performance. RESULTS: When directly prompted to predict diagnoses, the best LLM by F1 statistic (gpt-4o) had sensitivity of 76.7 % and PPV of 40.4 %. When making use of the refined decision trees, PPV was significantly increased (65.3 %) without a significant reduction in sensitivity (70.9 %). Across all experiments, the use of the decision trees statistically significantly increased the PPV, significantly increased the F1 statistic in 5/6 experiments, and significantly reduced sensitivity in 4/6 experiments. DISCUSSION: When used to predict psychiatric diagnoses from case vignettes, direct prompting of the LLMs yielded most true positive diagnoses but had significant overdiagnosis. Integrating expert-derived reasoning into the process using decision trees improved LLM performance (as measured by F1 statistic), primarily by suppressing overdiagnosis with a lower-magnitude negative impact on sensitivity. This suggests that the integration of clinical expert-derived reasoning could improve the performance of LLM-based tools in the behavioral health setting.

Humans↗
PubMed2026

An embedding-based framework enables statistical testing of gene-set function hypotheses inferred by large language models.

Emerging large language models (LLMs) can infer gene functions directly from gene lists, enabling hypothesis generation without predefined gene sets. However, these LLM-derived predictions are qualitative, and principled statistical validation is lacking. Here, we develop an embedding-based statistical framework that transforms gene and function descriptions into vector representations, enabling statistical testing of gene-gene and gene-function relationships and quantitative prioritization of de novo functional hypotheses inferred by LLMs. We benchmark seven state-of-the-art embedding models using curated and retrieval-augmented literature-derived gene descriptions across diverse biological contexts. OpenAI's text-embedding-3-large and Google's gemini-embedding-001 perform best, capturing gene-gene functional relationships in 88.7-92.5% of Gene Ontology biological processes and approximately 98.6% of canonical pathways. In gene-function association analyses, these models achieve high sensitivity (95.2-98.4%) and specificity (72.7-84.3%). Through contamination analysis and evaluation using experimentally informed protein assembly gene sets, our framework distinguishes biologically meaningful LLM-inferred hypotheses from noise, outperforming confidence-based inference and conventional enrichment analysis. We further develop the open-source R package DEGEmbedR and demonstrate its utility for interpreting a drug perturbation-derived differentially expressed gene (DEG) signature lacking significant conventional enrichment results. Together, these results establish LLM-derived embeddings as a quantitative foundation for functional genomics and the statistical validation of LLM-based gene function inference.

Large Language Models↗
PubMed1991

Detection of differential sensitivity to 5-fluorouracil in Ehrlich ascites tumour cells by 19F NMR spectroscopy.

Quantitative analysis of extracts from two Ehrlich ascites tumour cell lines (Lettre cells) by 19F NMR in vitro demonstrated that one Lettre cell line (designated LLM) metabolized 30-50% less 5-fluorouracil to 5-fluoronucleotides when compared to the other cell line (designated LHM). HPLC analysis of these cellular extracts showed a significant decrease in the concentration of the cytotoxic nucleotide 5-fluorouridine triphosphate in LLM cells compared to LHM cells. No major differences could be observed in the 31P and 1H NMR spectra of the two cell lines. Growth inhibition studies in vitro demonstrated that LLM cells were less sensitive to 5-fluorouracil than LHM cells. These results are consistent with the hypothesis that 19F NMR visible levels of 5-fluoronucleotides can predict the cytotoxicity of the anti-cancer drug 5-fluorouracil.

Adenosine Diphosphate↗
PubMed1980

Local leukocyte mobilization in irradiated or cyclophosphamide-treated rats.

Sprague-Dawley rats made neutropenic by 60Co irradiation or cyclophosphamide treatment retained a limited capacity for mounting a local leukocyte mobilization (LLM) response. Rats irradiated with 700 rad 60Co lacked an LLM. Rats treated with 100 mg/kg cyclophosphamide showed no LLM following an initial low response when assayed originally.

Agranulocytosis↗
PubMed2025

Automating candidate gene prioritization with large language models: from naive scoring to literature-grounded validation.

MOTIVATION: Identifying promising therapeutic targets from thousands of genes in transcriptomic studies remains a major bottleneck in biomedical research. While large language models (LLMs) show potential for gene prioritization, they suffer from hallucination and lack systematic validation against expert knowledge. RESULTS: The framework identified 609 sepsis-relevant genes with >94% filtering efficiency, demonstrating strong enrichment for inflammatory pathways including TNF-&#x3b1; signaling, complement activation, and interferon responses. Literature validation yielded 30 ultra-high confidence therapeutic candidates, including both established sepsis genes (IL10, TREM1, S100A9, NLRP3) and novel targets warranting investigation. Benchmark validation against expert-curated databases achieved 71.2% recall, with systematic correlation between computational confidence and evidence quality. The final candidate set balanced discovery (11 novel genes) with validation (19 known genes), maintaining biological coherence throughout the filtering process. This framework demonstrates that rigorous methodology can transform unreliable LLM outputs into systematically validated biological insights. By combining computational efficiency with literature grounding, the approach provides a practical tool for prioritizing experimental validation efforts. The modular design enables adaptation to other diseases through knowledge base substitution, offering a systematic approach to literature-guided biomarker discovery. AVAILABILITY AND IMPLEMENTATION: We developed a two-stage computational framework that combines LLM-based screening with literature validation for systematic gene prioritization. Starting with 10&#xa0;824 genes from the BloodGen3 repertoire, we applied multi-criteria evaluation for sepsis relevance, followed by retrieval-augmented generation using 6346 curated sepsis publications. A novel faithfulness evaluation system verified that LLM predictions aligned with retrieved literature evidence. Source code and implementation details are available at https://github.com/taushifkhan/llm-geneprioritization-framework, vector database at https://doi.org/10.5281/zenodo.15802241, and Interactive demonstration at https://llm-geneprioritization.streamlit.app/.

Humans↗