Lim Shan Chun, Itharajula Manoj, Møller Mads Harder, Sunil Rohan Shawn, Fo Kevin, Chuah Yu Song, Foo Herman, Davey Emilia Emmanuelle, Fullwood Melissa, Thibault Guillaume, Mutwil Marek
School of Biological Sciences, Nanyang Technological University, 60 Nanyang Drive, Singapore 637551, Singapore.
Section of Plant Biochemistry, Department of Plant and Environmental Sciences, University of Copenhagen, Thorvaldsensvej 40, Frederiksberg, Copenhagen 1871, Denmark.
Plant Cell. 2025 Jul 1;37(7). doi: 10.1093/plcell/koaf169.
One of the main quests in plant biology is understanding how gene products and metabolites work together to form complex networks that drive plant development and responses to environmental stimuli. However, the ever-growing volume and diversity of scientific literature make it increasingly challenging to stay current with the latest advances in functional genetics studies. Here, we tackled this challenge by deploying the text-mining capacities of large language models to process over 71,000 plant biology abstracts. Our approach presents nearly 5 million functional relationships between 2.4 million biological entities-genes or gene products, metabolites, tissues, and others-with a high accuracy of over 85%. We encapsulated these findings in the user-friendly database PlantConnectome and demonstrated its diverse utility by providing insights into gene regulatory networks, protein-protein interactions, and stress responses. We believe this innovative use of artificial intelligence (AI) in the life sciences will allow plant scientists to keep up to date with the rapidly growing corpus of scientific literature. PlantConnectome is available at https://plant.connectome.tools/.
植物生物学的主要探索之一是了解基因产物和代谢物如何共同作用，形成驱动植物发育和对环境刺激作出反应的复杂网络。然而，科学文献数量的不断增长及其多样性，使得紧跟功能遗传学研究的最新进展变得越来越具有挑战性。在此，我们通过运用大语言模型的文本挖掘能力来处理超过71000篇植物生物学摘要，应对了这一挑战。我们的方法呈现了240万个生物实体（基因或基因产物、代谢物、组织等）之间近500万个功能关系，准确率超过85%。我们将这些发现整合到用户友好的数据库PlantConnectome中，并通过深入了解基因调控网络、蛋白质-蛋白质相互作用和应激反应，展示了其多样的用途。我们相信，人工智能在生命科学中的这种创新应用将使植物科学家能够跟上迅速增长的科学文献体量。可通过https://plant.connectome.tools/访问PlantConnectome。
