Developing a new generation of tools for predicting novel AMR mutation profiles using generative AI

治疗传染病的药物改变了人类和动物的健康，拯救了数百万人的生命。然而，它们的广泛使用和误用导致了抗菌素耐药性(AMR)的出现，这对公共卫生和畜牧业构成了潜在的灾难性威胁。主要途径之一，也是该项目的一个重点，是通过在基因组区域进行单点突变来编码蛋白质，并导致蛋白质的氨基酸序列发生变化。这些类型的突变称为单核苷酸多态(SNPs)。基因组测序的进步意味着现在有来自一系列病原体的大量序列集合，在这些序列中已经发现了SNPs，并可能与耐药性有关。该项目旨在利用这些丰富的数据，结合最近在准确建模蛋白质结构方面取得的进展，开发一种新的基于人工智能的工具来预测可能导致耐药性的SNPs的影响，该工具尚未被观察到。通过模拟病原体如何突变以避免药物的影响，我们可以更好地预测感染对特定药物的反应，并可能设计出临床使用时间更长的药物。除了直接受益于那些致力于开发下一代药物的人，它还使那些管理处方常规和监测的人受益，在新出现的耐药性在人群中广泛传播之前发现可以采取行动的新出现的耐药性。该项目汇集了来自澳大利亚昆士兰大学和伦敦卫生与热带医学院(LSHTM，UK)的一组国际专家，他们在人工智能、耐药性和细菌病原体基因组学方面拥有互补的专业知识。该项目有几个关键目标：目标1：开发一个基于自然语言处理(NLP)的人工智能工具，用于预测引起耐药性的SNPs，训练来自与耐药性相关的突变的大量病原体基因组数据的特征。目标2：验证并将新开发的方法应用于特定病原体，包括伤寒沙门氏菌和肺炎克雷伯菌(WHO优先病原体)，为现实世界的验证提供机会，并提供深入了解耐药性机制的能力。目标3：通过两个英国主导的研讨会，将人工智能应用于AMR的知识交流。这将加强协作网络，根据用户需求建立人工智能工具的设计标准，并提供一条将工具转化为现实世界使用的途径。此外，英国和澳大利亚团队之间的研究人员交流将增强两个团队的能力和能力。该项目设想了一种人工智能支持的解决方案，以帮助先发制人地应对耐药性突变的影响，解决抗击AMR日益增长的威胁的迫切需求。经过验证的新计算工具将有助于开发更好的药物，并与补充技术一起，帮助决定药物治疗方案和耐药性监测。它将促成以英国为首的国际合作伙伴关系，使参与其中的组织走在该领域研究的前沿。

Drugs against infectious diseases have transformed human and animal health and saved millions of lives. Nevertheless, their widespread use and misuse has led to the emergence of antimicrobial resistance (AMR) that poses a potentially catastrophic threat to public health and animal husbandry.There a several routes by which a pathogen can become resistant to a drug. One of the principal routes, and a focus of this project, is by single point mutations in genomic regions that code for proteins and result in a change in the protein's sequence of amino acids. These types of mutations are called Single Nucleotide Polymorphisms (SNPs). Advances in genome sequencing means there are now large collections of sequences from a range of pathogens where SNPs have been identified and can be associated with drug resistance. This project aims to capitalise on this wealth of data, combined with the recent advances made to accurately model protein structures, to develop a new AI-based tool to predict the effect of SNPs that could lead to resistance and have yet to be observed.By modelling how pathogens mutate to avoid the effect of drugs, we can better predict how infections will respond to specific drugs and may be able to design drugs that have longer clinical use. As well as directly benefiting those working to develop the next generation of drugs, it also benefits those managing prescribing routines and in surveillance, identifying new emerging resistance that can be acted on before it becomes widespread within a population.The project brings together a group of international experts from the University of Queensland (Australia) and the London School of Hygiene & Tropical Medicine (LSHTM, UK) who have complementary expertise in AI, drug resistance and bacterial pathogen genomics. The project has several key objectives:Objective 1: Develop a Natural Language Processing (NLP)-based AI tool for predicting SNPs causing resistance trained on features derived from the large collections of pathogen genome data where mutations associated with drug resistance have been identified.Objective 2: Validate and apply the newly developed methodologies to specific pathogens including Salmonella Typhiand Klebsiella pneumoniae (WHO priority pathogens) that gives opportunity for real-world validation and the ability to give insights into resistance mechanisms.Objective 3: Knowledge Exchange of AI applied to AMR though two UK-led workshops. This will enhance the collaborative network, establish design criteria for the AI tool based on user needs, and provide a pathway to translating the tools into real-world use. In addition, exchanges of researchers between the UK and Australian groups will enhance capacity and capabilities of both teams.This project envisions an AI-powered solution to help pre-empt the impact of drug resistance mutations, addressing the urgent need to combat the growing threat of AMR. The validated new computational tools will help in developing better drugs and, in conjunction with complementary technologies, aid in deciding drug treatment regimens and in resistance surveillance. It will enable a UK-led international partnership that will place the groups involved at the forefront of research in this field.