Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “agentic AI”

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

Reimagining research papers as interactive and reliable AI agents.

Here we introduce Paper2Agent, an automated framework that converts research papers into artificial intelligence (AI) agents. Paper2Agent transforms research output from passive artefacts into active systems that accelerate use and discovery. Conventional research papers require readers to understand and adapt the paper's code, data and methods to their work, creating barriers to dissemination and reuse. Paper2Agent addresses this challenge by converting a paper into an AI agent that functions as a virtual corresponding author, exposing its manuscript, supplementary materials, datasets, code and workflows as active, agent-native knowledge rather than static text. It analyses the paper and codebase using multiple agents to construct a model context protocol (MCP) server, then generates and runs tests to refine and increase robustness of the MCP. These paper MCPs can be connected to a chat agent (such as Claude Code) to carry out complex scientific queries through natural language while invoking tools and workflows from the paper. We demonstrate Paper2Agent's effectiveness through case studies. Paper2Agent created an agent that leveraged AlphaGenome1 to interpret genomic variants and agents based on Scanpy2 and TISSUE (transcript imputation with spatial single-cell uncertainty estimation)3 to conduct single-cell and spatial transcriptomics analyses. We validate that these agents reproduce the results of the original papers and carry out novel user queries. Paper2Agent created multiple agents that collaborate to prioritize a causal gene for psoriasis. By turning static papers into interactive AI agents, Paper2Agent introduces a paradigm for knowledge dissemination and a collaborative ecosystem of AI co-scientists.

Journal Article↗
PubMed2002

Evaluation of a new ultrasound contrast agent (AI-700) using two-dimensional and three-dimensional imaging during acute ischemia.

BACKGROUND: A new intravenous contrast agent, AI-700, was evaluated to determine whether a bolus injection could be used to detect myocardial perfusion abnormalities during acute ischemia by using 2-dimensional (2D) and 3-dimensional (3D) myocardial contrast echocardiography. METHODS: 2D MCE was performed in 14 closed-chest dogs during coronary occlusion by using both continuous and triggered gray scale harmonic imaging and triggered power Doppler imaging. 3D MCE (open-chest) and nuclear perfusion imaging were performed in 10 of the 14 dogs. Postmortem triphenyl tetrazolium chloride (TTC) staining was performed to verify infarction. RESULTS: Thirteen of the 14 dogs had infarct by TTC; all 10 that had nuclear imaging showed a perfusion defect. Of the 13 dogs that had infarction, perfusion defects were detected in all (13 of 13) by gray scale harmonic imaging (sensitivity = 100%), and in 11 of 13 by power Doppler imaging (sensitivity = 85%). All 10 dogs that had nuclear imaging showed perfusion defects by gray scale harmonic imaging (sensitivity = 100%) and 8 of 10 by power Doppler imaging (sensitivity = 80%). The perfusion defect size, derived from 3D imaging (25% +/- 12%) correlated well with that from nuclear imaging (24% +/- 12%) (y = 0.9x + 3.8, r = 0.96, mean difference = 1.3% +/- 2.6%). The perfusion defect mass by 3D (22 +/- 14 g) also correlated well with the infarct mass by TTC staining (24 +/- 16 g) (y = 0.8x + 2.9, r = 0.89, P <.001, mean difference = -2.8 +/- 7.6 g). CONCLUSION: After a single bolus of AI-700, both 2D and 3D MCE could accurately detect perfusion defects representing the area at risk of infarction during acute ischemia compared with nuclear imaging and predicted the size of infarction as verified by TTC staining.

Animals↗
PubMed2026

Agentic AI for Spatial Omics.

This highlight summarises recent advances in agentic artificial intelligence (AI) systems for spatial omics analysis. These systems are compared along two central tensions: autonomy versus accountability, and adaptability versus reproducibility. We argue that progress will depend not on maximising automation, but on defining where autonomy is appropriate.

Artificial Intelligence↗
PubMed2026

Trustworthy Agentic AI in Bioinformatics: From Workflow Automation to Traceable and Validated Biological Inference.

Agentic artificial intelligence is extending bioinformatics beyond conversational assistance by enabling systems to select tools, execute code, revise analytical plans, and interpret biological data. These capabilities may accelerate research, but they also redistribute decisions that determine whether biological conclusions are valid. We conducted a targeted, structured PubMed search in July 2026 and identified 11 peer-reviewed agentic bioinformatics systems for descriptive review based on predefined eligibility criteria for analytical decision-making, tool or code execution, iterative evaluation, or coordinated agent activity. The evidence base covered single-cell transcriptomics, microbial genomics, cancer genomics, and omics applications, together with methodological literature on reproducibility and biological validation. We examined how current systems report delegated authority, provenance, validation, evidence, abstention, and human oversight. Existing platforms implement safeguards such as sandboxed execution, restricted commands, interaction logs, evidence identifiers, automated checks, critic agents, quality scores, and expert assessment. However, published reports rarely provide a connected account linking the original biological question to samples, reference resources, analytical decisions, computational actions, statistical results, supporting evidence, validation outcomes, and final claims. We distinguish inherited bioinformatics errors, errors amplified through autonomous action, and emergent failures arising from memory, retrieval, tool interaction, or agent coordination. We further propose a multidimensional decision-rights profile, consequence-sensitive validation gates, and a claim-to-evidence provenance architecture organized through the Traceable History of Research Evidence, Agent Actions, and Decisions in Bioinformatics (THREAD-Bio) framework. Illustrative cases show that technically successful execution may still support misleading inference. Trustworthy agentic bioinformatics therefore requires claims to remain reconstructible, challengeable, validated, and proportionate to the evidence.

accountable autonomy↗
PubMed2001

Stereoselective synthesis of pseudopeptide microbial agent AI-77-B.

[structure: see text]. An efficient and highly stereoselective synthesis of the gastroprotective natural product AI-77-B is described. The stereocenters of the hydroxy amino acid moiety were generated by an ester-derived titanium-enolate-mediated syn-aldol reaction, a Curtius rearrangement, and application of Dondoni's aldehyde homologation. Condensation with the dihydroisocoumarin fragment and subsequent deprotecting transformations furnished optically active AI-77-B.

Anti-Bacterial Agents↗
PubMed2026

scBaseCount: An AI agent-curated, standardized, auto-updated single-cell data repository.

Single-cell RNA sequencing has transformed cell biology by enabling precise transcriptomic measurements of individual cells. The Sequence Read Archive (SRA) is the largest public repository of sequencing reads, yet much of it remains underutilized due to unstandardized metadata. Here, we introduce scBaseCount, a database that leverages an AI agent to automate discovery and metadata extraction and standardize data processing. Built by mining all 10x Genomics datasets, scBaseCount is the largest public repository of single-cell gene expression data, comprising over 502 million cells across 27 organisms and 75 tissues. It offers an unbiased view of the data landscape within the SRA and enables the training of more performant computational models through access to broader phenotypic diversity. Uniform processing enables measurement of both intronic and exonic reads and non-coding gene expression and improves alignment across experiments. Moreover, scBaseCount provides a blueprint for how AI can be leveraged to autonomously curate biological data repositories.

Single-Cell Analysis↗
PubMed2006

A cognitive approach to game usability and design: mental model development in novice real-time strategy gamers.

We developed a technique to observe and characterize a novice real-time-strategy (RTS) player's mental model as it shifts with experience. We then tested this technique using an off-the-shelf RTS game, EA Games Generals. Norman defined mental models as, "an internal representation of a target system that provides predictive and explanatory power to the operator." In the case of RTS games, the operator is the player and the target system is expressed by the relationships within the game. We studied five novice participants in laboratory-controlled conditions playing a RTS game. They played Command and Conquer Generals for 2 h per day over the course of 5 days. A mental model analysis was generated using player dissimilarity-ratings of the game's artificial intelligence (AI) agents analyzed using multidimensional scaling (MDS) statistical methods. We hypothesized that novices would begin with an impoverished model based on the visible physical characteristics of the game system. As they gained experience and insight, their mental models would shift and accommodate the functional characteristics of the AI agents. We found that all five of the novice participants began with the predicted physical-based mental model. However, while their models did qualitatively shift with experience, they did not necessarily change to the predicted functional-based model. This research presents an opportunity for the design of games that are guided by shifts in a player's mental model as opposed to the typical progression through successive performance levels.

Cognition↗
PubMed2026

ELISA (Embedding-Linked Interactive Single-cell Agent): an interpretable hybrid generative Artificial Intelligence agent for expression-grounded discovery in single-cell genomics.

Translating single-cell RNA sequencing (scRNA-seq) data into mechanistic biological hypotheses remains a critical bottleneck, as agentic AI systems lack direct access to transcriptomic representations while expression foundation models remain opaque to natural language. Here, we introduce ELISA (Embedding-Linked Interactive Single-cell Agent), an interpretable framework that unifies single-cell generative pretrained transformer expression embeddings with biomedical bidirectional encoder representations from transformers-based semantic retrieval and large-language model (LLM)-mediated interpretation for interactive single-cell discovery. An automatic query classifier routes inputs to gene marker scoring, semantic matching, or reciprocal rank fusion pipelines depending on whether the query is a gene signature, natural language concept, or mixture of both. Integrated analytical modules perform pathway activity scoring across 60+ gene sets, ligand-receptor interaction prediction using 280+ curated pairs, condition-aware comparative analysis, and cell-type proportion estimation, all operating directly on embedded data without access to the original count matrix. Benchmarked across six diverse scRNA-seq datasets spanning inflammatory lung disease, pediatric and adult cancers, organoid models, healthy tissue, and neurodevelopment, ELISA significantly outperforms CellWhisperer, a classical lexical retriever (BM25), and a random baseline in cell type retrieval (combined permutation test, $p < 2\times 10^{-5}$ for each), with particularly large gains on gene-signature queries (Cohen's $d = 5.98$ for mean reciprocal rank). ELISA replicates published biological findings (mean composite score 0.88), and generates candidate hypotheses through grounded LLM reasoning, bridging the gap between transcriptomic data exploration and biological discovery.

Generative Artificial Intelligence↗
PubMedDate not supplied

Postmortem studies of the effect of anti-inflammatory drugs on Alzheimer-type pathology and associated inflammation.

Examining postmortem tissue is the most direct way of evaluating the effect of antemortem drug use on the pathological processes believed to be important in Alzheimer's disease (AD). A small number of studies have recently been published in which data from human autopsy tissue and animal models provides important insight into the mechanisms by which anti-inflammatory (AI) agents may protect against AD. These indicate that certain classes of AI drugs may be capable of reducing the chronic inflammation which is consistently seen in AD brain tissue. In addition, a recent study using a transgenic mouse model of AD, suggests that AI therapy may also influence the accumulation of senile plaques and dystrophic neurites. The results of these and future postmortem studies will be invaluable in the development of optimum treatment strategies.

Alzheimer Disease↗
PubMed1985

Effects of anti-inflammatory drugs on catabolin-induced cartilage destruction in vitro.

Catabolin/Interleukin 1 (CAT-IL-1) comprises a family of acidic proteins produced by stimulated monocytes and inflamed synovium which cause the release of proteoglycans and collagen from living cartilage in vitro. The possible relevance of these proteins as mediators of joint destruction in animal models of joint damage and in arthritic disease in man has yet to be determined. To establish if conventional or novel anti-inflammatory (AI) agents could modify the actions of CAT-IL-1 and thus be useful for understanding the mode of action of these components, the effects were determined of a range of AI compounds on the release by CAT-IL-1 of proteoglycans (PG's) from bovine nasal cartilage. In vitro, conventional non-steroidal anti-inflammatory (NSAI) compounds were, with the exception of 10-100 microM hydroxychloroquine, weak (e.g. diclofenac, indomethacin, fenclofenac, effective at concentrations of 100-500 microM), or with some drugs actually ineffective, as inhibitors of CAT-IL-1-mediated PG release. Antioxidants were ineffective as inhibitors of this PG release. Drugs which modify gene transcription were potent inhibitors of CAT-IL-1-induced PG release, emphasizing the importance of gene expression in the actions of these mediators. The importance of prostaglandin metabolism on CAT-IL-1 actions and the in vivo effects of AI are discussed.

Animals↗
PubMed2001

Design and synthesis of 4,5-diphenyl-4-isoxazolines: novel inhibitors of cyclooxygenase-2 with analgesic and antiinflammatory activity.

4,5-Diphenyl-4-isoxazolines (13a-k) possessing a variety of substituents (H, F, MeS, MeSO2) at the para position of one of the phenyl rings were synthesized for evaluation as analgesic and selective cyclooxygenase-2 (COX-2) inhibitory antiinflammatory (AI) agents. Although the 4,5-phenyl-4-isoxazolines (13a-d,f), which do not have a C-3 Me substituent, exhibited potent analgesic and AI activities, those compounds evaluated (13a, 13b, 13h, and 13k) were not selective inhibitors of COX-2. In contrast, 2,3-dimethyl-5-(4-methylsulfonylphenyl)-4-phenyl-4-isoxazoline (13j) exhibited excellent analgesic and AI activities, and it was a potent and selective COX-2 inhibitor (COX-1, IC(50) = 258 microM; COX-2, IC(50) = 0.004 microM). A related compound 13k having a F substituent at the para position of the 4-phenyl ring was also a selective (SI = 3162) but less potent (IC(50) = 0.0316 microM) inhibitor of COX-2 than 13j. A molecular modeling (docking study) for 13j showed that the S atom of the MeSO2 substituent is positioned about 6.46 A inside the entrance to the COX-2 secondary pocket (Val(523)) and that a C-3 Me (13j, 13k) central isoxazoline ring substituent is crucial to selective inhibition of COX-2 for this class of compounds.

Abdominal Muscles↗
PubMed1998

Preventive pharmacologic therapy among asthmatics: five years after publication of guidelines.

BACKGROUND: Guidelines from the National Heart, Lung, and Blood Institute first published in 1991 have recommended anti-inflammatory (AI) agents as a first-line therapy and the bronchodilator as an acute reliever of symptoms. OBJECTIVE: To examine the current usage of anti-inflammatory steroids (inhaled corticosteroids, Cromolyn, systemic steroids) and bronchodilators and compare them with the national guidelines. The relationship between preventive AI usage and the characteristics of the asthma patients and their providers was also examined. METHODS: Cross-sectional survey data linked with 6-month pharmacy claims of asthmatic members at an HMO in California. RESULTS: AI usage increased with current severity (mild, 36.9%; moderate, 47.3%; and severe, 56.8%), though a large percentage are not receiving this emphasized treatment. Bronchodilators were used at a higher rate and 24% of asthmatics relied solely on bronchodilators. Use of bronchodilators without AI (BWAI) was present at all severity levels (mild, 19.5%; moderate, 24.6%; and severe, 24.7%). Advancing age, increasing severity, care by an asthma specialist, and not smoking increased the likelihood of using AIs. Increasing severity, longer duration of asthma, smoking, younger age group, care by a generalist, and no chronic bronchitis increased the likelihood of BWAI. CONCLUSIONS: These results suggest that there is a low level of AI usage despite emphasis in guidelines. Current asthma management in a community-based setting depicts a significant underutilization of long-term control agents and, conversely, an overutilization of symptom relief agents compared with guidelines published 5 years ago. Actively involving patients in the guideline dissemination process, rather than just the medical community, may increase preventive medication usage.

Adolescent↗
PubMed2025

Out-of-the-box bioinformatics capabilities of large language models (LLMs).

Large Language Models (LLMs), AI agents and co-scientists promise to accelerate scientific discovery across fields ranging from chemistry to biology. Bioinformatics- the analysis of DNA, RNA and protein sequences plays a crucial role in biological research and is especially amenable to AI-driven automation given its computational nature. Here, we assess the bioinformatics capabilities of three popular general-purpose LLMs on a set of tasks covering basic analytical questions that include code writing and multi-step reasoning in the domain. Utilizing questions from Rosalind, a bioinformatics educational platform, we compare the performance of the LLMs vs. humans on 104 questions undertaken by 110 to 68,760 individuals globally. GPT-3.5 provided correct answers for 59/104 (58%) questions, while Llama-3-70B and GPT-4o answered 49/104 (47%) correctly. GPT-3.5 was the best performing in most categories, followed by Llama-3-70B and then GPT-4o. 71% of the questions were correctly answered by at least one LLM. The best performing categories included DNA analysis, while the worst performing were sequence alignment/comparative genomics and genome assembly. Overall, LLMs performance mirrored that of humans with lower performance in tasks in which humans had low performance and vice versa. However, LLMs also failed in some instances where most humans were correct and, in a few cases, LLMs excelled where most humans failed. To the best of our knowledge, this presents the first assessment of general purpose LLMs on basic bioinformatics tasks in distinct areas relative to the performance of hundreds to thousands of humans. LLMs provide correct answers to several questions that require use of biological knowledge, reasoning, statistical analysis and computer code.

Journal Article↗
PubMed1994

Clinical study of biological response modifiers as maintenance therapy for hepatocellular carcinoma.

We conducted a randomized, controlled trial comparing 5-fluorouracil (5-FU) with or without biological response modifiers (BRMs) as a maintenance therapy for hepatocellular carcinoma (HCC) after treatment with percutaneous ethanol injection (PEI), transcatheter arterial embolization (TAE) or arterial infusion of antitumor agents (AI). A total of 58 cases of HCC were classified into 4 groups as follows: group I, PSK with 5-FU (n = 15); group II, lentinan with 5-FU (n = 15); group III, OK-432 with 5-FU (n = 12); and group IV, 5-FU alone as the control (n = 16). The mean survival time, mortality rate, time to progression, and T4/T8 ratio of lymphocytes in the peripheral blood were compared among the four groups. There was no significant difference in the background factors among the groups. In group I, the T4/T8 ratio of lymphocytes was reduced after the therapy. No significant difference was found among the groups in terms of the mean survival time, mortality rate, or time to progression. PEI for initial therapy was superior to the other therapies in terms of the mean survival time and mortality rate. These results suggest that the addition of BRM to maintenance therapy with 5-FU exerts no prognostic benefit on HCC patients treated with PEI, TAE, or AI.

Aged↗
PubMed2004

Design and synthesis of acyclic triaryl (Z)-olefins: a novel class of cyclooxygenase-2 (COX-2) inhibitors.

A group of acyclic 2-alkyl-1,1-diphenyl-2-(4-methylsulfonylphenyl)ethenes was designed for evaluation as selective cyclooxygenase-2 (COX-2) inhibitors. In vitro COX-1 and COX-2 isozyme inhibition structure-activity studies identified 1,1-diphenyl-2-(4-methylsulfonylphenyl)hex-1-ene as a highly potent (IC(50) = 0.014 microM), and an extremely selective [COX-2 selectivity index (SI) > 7142], COX-2 inhibitor that showed superior anti-inflammatory (AI) activity (ID(50) = 2.5 mg/kg) relative to celecoxib (ID(50) = 10.8 mg/kg). This initial study was extended to include the design of a structurally related group of acyclic triaryl (Z)-olefins possessing an acetoxy (OAc) substituent at the para-position of the C-1 phenyl ring that is cis to a C-2 4-methylsulfonylphenyl substituent. COX-1 and COX-2 inhibition studies showed that (Z)-1-(4-acetoxyphenyl)-1-phenyl-2-(4-methylsulfonylphenyl)but-1-ene [(Z)-13b] is a potent (COX-1 IC(50) = 2.4 microM; COX-2 IC(50) = 0.03 microM), and selective (COX-2 SI = 81), COX-2 inhibitor which is a potent AI agent (ID(50) = 4.1mg/kg) with equipotent analgesic activity to celecoxib. A molecular modeling (docking) study showed that the SO(2)Me substituent of (Z)-13b inserts deep inside the 2 degrees -pocket of the COX-2 active site, where one of the O-atoms of SO(2) group undergoes a H-bonding interaction with Phe(518). The p-OAc substituent on the C-1 phenyl ring is oriented in a hydrophobic pocket comprised of Met(522), Gly(526), Trp(387), Tyr(348), and Tyr(385), and the C-2 ethyl substituent is oriented towards the mouth of the COX-2 channel in the vicinity of amino acid residues Arg(120), Leu(531), and Val(349). Structure-activity data acquired indicate that a (Z)-olefin having cis C-1 4-acetoxyphenyl (phenyl) and C-2 4-methylsulfonylphenyl substituents, and a C-1 phenyl substituent in conjunction with either a C-2 hydrogen or short alkyl substituent provides a novel template to design acyclic olefinic COX-2 inhibitors that, like aspirin, have the potential to acetylate COX-2.

Alkenes↗
PubMed2026

Biological Foundation Models for Complex Disease Research and Clinical Translation.

Complex diseases, including cancer, rare genetic disorders, neurodevelopmental and psychiatric conditions, and neurodegenerative diseases, arise from interactions among genetic variation, gene regulation, and cellular states that are difficult to capture using a single data type or biological scale. Biological foundation models address this challenge by treating nucleotides and genes as tokens and learning representations that can be transferred to downstream biomedical and clinical tasks. In this review, we examine two major model classes, genomic sequence foundation models and cell foundation models, and compare their tokenization strategies, model architectures, pretraining objectives, and adaptation methods. We summarize their emerging applications in regulatory variant interpretation, disease-associated cell-state analysis, drug-response prediction, and therapeutic target discovery across complex diseases. We distinguish applications supported by experimental or retrospective validation from those that remain primarily computational or conceptual. We further discuss key challenges to clinical translation, including multimodal data integration, model interpretability, benchmarking, patient-specific prediction, and privacy protection. We highlight future opportunities to integrate biological foundation models with emerging frameworks of medical digital twins, agentic AI, and federated learning. By linking model design to translational goals, this review provides a practical framework for evaluating biological foundation models and their readiness for complex disease research and clinical use.

biological foundation model↗
PubMed2025

AI-HOPE: an AI-driven conversational agent for enhanced clinical and genomic data integration in precision medicine research.

MOTIVATION: The growing complexity of clinical cancer research has fueled a surge in demand for automated bioinformatics tools capable of integrating clinical and genomic data to accelerate discovery efforts. RESULTS: We present the Artificial Intelligence Agent for High-Optimization and Precision Medicine (AI-HOPE), an AI-driven system that enables domain experts to conduct integrative data analyses through natural language interactions. Powered by Large Language Models, AI-HOPE interprets user instructions, converts them into executable code, and autonomously analyzes locally stored data. It supports flexible association studies, subset comparisons, clinical prevalence assessments and survival analyses. In addition, AI-HOPE enables global variable scans to identify features significantly associated with a user-defined outcome, making a powerful and intuitive tool for advancing precision medicine research. Importantly, its closed-system design prevents clinical data leakage. To demonstrate its utility, AI-HOPE was applied to The Cancer Genome Atlas data to address two clinical questions. First, it identified significant enrichment of TP53 mutations in late-stage colorectal cancer compared to early-stage cases. Second, it uncovered a strong association between KRAS mutations and poorer progression-free survival in FOLFOX-treated patients. These findings align with established literature and demonstrate AI-HOPE's ability to generate meaningful insights independently, without prior assumptions. By removing programming barriers and simplifying complex analyses, AI-HOPE bridges the gap between data complexity and research needs. With its scalable and adaptable framework, AI-HOPE has the potential to support diverse biomedical research fields, driving innovation and efficiency in translational studies. AVAILABILITY AND IMPLEMENTATION: The AI-HOPE software and demonstration data is available at https://github.com/Velazquez-Villarreal-Lab/AI-HOPE.

Precision Medicine↗