Large-language models facilitate discovery of the molecular signatures regulating sleep and activity

Nature Communications, Год журнала: 2024, Номер 15(1)

Опубликована: Май 1, 2024

Abstract Sleep, locomotor and social activities are essential animal behaviors, but their reciprocal relationships underlying mechanisms remain poorly understood. Here, we elicit information from a cutting-edge large-language model (LLM), generative pre-trained transformer (GPT) 3.5, which interprets 10.2–13.8% of Drosophila genes known to regulate the 3 behaviors. We develop an instrument for simultaneous video tracking multiple moving objects, conduct genome-wide screen. have identified 758 fly that sleep activities, including mre11 regulates only in presence conspecifics, NELF-B regardless whether conspecifics present. Based on LLM-reasoning, educated signal web is modeled understanding potential between its components, presenting comprehensive molecular signatures control sleep, activities. This LLM-aided strategy may also be helpful addressing other complex scientific questions.

Язык: Английский

---

Overview and Evolution of Insect Fibroin Heavy Chain (FibH)

International Journal of Molecular Sciences, Год журнала: 2024, Номер 25(13), С. 7179 - 7179

Опубликована: Июнь 29, 2024

The

Язык: Английский

---

Understanding the Role of Large Language Models in Shaping Gen-Z Tourist Expectations

Advances in computational intelligence and robotics book series, Год журнала: 2025, Номер unknown, С. 467 - 494

Опубликована: Март 12, 2025

This study explores the role of Large Language Models in shaping travel planning behaviours among students Indian educational institutions. With increasing reliance on AI technologies, LLMs, such as OpenAI's GPT models, are becoming prominent tools for travellers seeking personalised and efficient information. research investigates comparative reliability LLMs with traditional sources, social media, blogs, agencies. A total 143 were surveyed using a structured questionnaire facilitated by nine trained enumerators. The findings indicate that while media remains most frequently used source inspiration, valued their reliability, personalized recommendations, ability to enhance efficiency planning. offers valuable insights into how can reshape industry, giving theoretical contributions field tourism practical implications developers marketers technologies.

Язык: Английский

---

Artificial Intelligence in Central-Peripheral Interaction Organ Crosstalk: The Future of Drug Discovery and Clinical Trials

Pharmacological Research, Год журнала: 2025, Номер unknown, С. 107734 - 107734

Опубликована: Апрель 1, 2025

Drug discovery before the 20th century often focused on single genes, molecules, cells, or organs, failing to capture complexity of biological systems. The emergence protein-protein interaction network studies in 2001 marked a turning point and promoted holistic approach that considers human body as an interconnected system. This is particularly evident study bidirectional interactions between central nervous system (CNS) peripheral which are critical for understanding health disease. Understanding these complex requires integrating multi-scale, heterogeneous data from molecular organ levels, encompassing both omics (e.g., genomics, proteomics, microbiomics) non-omics imaging, clinical phenotypes). Artificial intelligence (AI), multi-modal models, has demonstrated significant potential analyzing CNS-peripheral by processing vast, datasets. Specifically, AI facilitates identification biomarkers, prediction therapeutic targets, simulation drug effects multi-organ systems, thereby paving way novel strategies. review highlights AI's transformative role research, focusing its applications unraveling disease mechanisms, discovering optimizing trials through patient stratification adaptive trial design.

Язык: Английский

---

Comprehensive benchmarking of large language models for RNA secondary structure prediction

Briefings in Bioinformatics, Год журнала: 2025, Номер 26(2)

Опубликована: Март 1, 2025

In recent years, inspired by the success of large language models (LLMs) for DNA and proteins, several LLMs RNA have also been developed. These take massive datasets as inputs learn, in a self-supervised way, how to represent each base with semantically rich numerical vector. This is done under hypothesis that obtaining high-quality representations can enhance data-costly downstream tasks, such fundamental secondary structure prediction problem. However, existing RNA-LLM not evaluated this task unified experimental setup. Since they are pretrained models, assessment their generalization capabilities on new structures crucial aspect. Nonetheless, has just partially addressed literature. work we present comprehensive comparative analysis recently proposed. We evaluate use these common deep learning architecture. The were assessed increasing difficulty benchmark datasets. Results showed two clearly outperform other revealed significant challenges low-homology scenarios. Moreover, study provide curated complexity setup scientific endeavor. Source code available repository: https://github.com/sinc-lab/rna-llm-folding/.

Язык: Английский

---

Evolution of gene regulatory networks in insects

Current Opinion in Insect Science, Год журнала: 2025, Номер 69, С. 101365 - 101365

Опубликована: Май 8, 2025

Язык: Английский

---

Upstream open reading frames dynamically modulate CLOCK protein translation to regulate circadian rhythms and sleep

PLoS Biology, Год журнала: 2025, Номер 23(5), С. e3003173 - e3003173

Опубликована: Май 12, 2025

The circadian rhythm is an evolutionarily conserved mechanism with translational regulation increasingly recognized as pivotal in its modulation. In this study, we found that upstream open reading frames (uORFs) are enriched Drosophila genes, particularly uORFs present core clock genes. We demonstrate evidence the of gene, Clock ( Clk ), rhythmically and substantially attenuate CLK protein translation , pronounced suppression occurring during daylight hours. Eliminating leads to increased levels day results a shortened cycle, along broad shift gene expression rhythms. Notably, uORF deletion also augments morning sleep by reducing dopaminergic activity. Beyond daily adjustments, play role modulating patterns response seasonal variations. Furthermore, act important regulator shape rhythmic vast array genes influence multifaceted physiological outcomes. Collectively, our research sheds light on intricate ways dynamically adjust downstream coding sequences acclimate environmental shifts.

Язык: Английский

---

Reporting guideline for the use of Generative Artificial intelligence tools in MEdical Research: the GAMER Statement

BMJ evidence-based medicine, Год журнала: 2025, Номер unknown, С. bmjebm - 113825

Опубликована: Май 13, 2025

Objectives Generative artificial intelligence (GAI) tools can enhance the quality and efficiency of medical research, but their improper use may result in plagiarism, academic fraud unreliable findings. Transparent reporting GAI is essential, yet existing guidelines from journals institutions are inconsistent, with no standardised principles. Design setting International online Delphi study. Participants experts medicine intelligence. Main outcome measures The primary measure consensus level expert panel on items inclusion criteria for GAMER (Rreporting guideline Artificial MEdical Research). Results development process included a scoping review, two rounds virtual meetings. 51 26 countries participated (44 survey). final checklist comprises nine items: general declaration, tool specifications, prompting techniques, tool’s role study, declaration new model(s) developed, intelligence-assisted sections manuscript, content verification, data privacy impact conclusions. Conclusion provides universal ensuring transparency, integrity quality.

Язык: Английский

---

Natural language processing in veterinary pathology: A commentary on opportunities, challenges, and future directions

Veterinary Pathology, Год журнала: 2025, Номер unknown

Опубликована: Май 26, 2025

Язык: Английский

---

GeneRAG: Enhancing Large Language Models with Gene-Related Task by Retrieval-Augmented Generation

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Июнь 28, 2024

Abstract Large Language Models (LLMs) like GPT-4 have revolutionized natural language processing and are used in gene analysis, but their knowledge is incomplete. Fine-tuning LLMs with external data costly resource-intensive. Retrieval-Augmented Generation (RAG) integrates relevant information dynamically. We introduce G ene RAG, a frame-work that enhances LLMs’ gene-related capabilities using RAG the Maximal Marginal Relevance (MMR) algorithm. Evaluations datasets from National Center for Biotechnology Information (NCBI) show outperforms GPT-3.5 GPT-4, 39% improvement answering questions, 43% performance increase cell type annotation, 0.25 decrease error rates interaction prediction. These results highlight RAG’s potential to bridge critical gap LLM more effective applications genetics.

Язык: Английский

---