Using Machine Learning to Anticipate Antimicrobial Resistance for Pathogen Surveillance and Therapeutic Stewardship

抗生素耐药性（AMR）是人类和动物健康面临的严重全球健康挑战。抗菌药物是用于对抗病原微生物（病毒，细菌和寄生虫）的药物。这些药物的广泛使用和滥用导致了抗药性的出现。微生物对药物产生耐药性并在药物暴露后存活的方式之一是经历遗传变化（突变）。迫切需要在特定病原体群体中固定之前识别和预测导致抗性的此类突变。技术进步导致了健康科学各个方面的快速，成本效益和广泛的数据收集。然而，为了充分利用这些信息丰富的来源，必须开发相应的计算技术，以提高我们对复杂生物问题的理解。通过机器学习（ML）实现的人工智能提供了一种强大的方法来加快我们对这些问题的理解。顾名思义，ML是一种让计算机从数据中学习和发现模式的方法。该项目旨在开发一个强大的自动化计算框架，以识别病原体中的新突变。了解这些突变的分子后果将提供深入了解日益失败的抗菌治疗。开发复杂的机器学习方法将加速识别检测和应对AMR的新方法。这将对病原体监测、援助管理和指导新干预措施的制定产生直接影响。这个多学科项目整合和提高数学，统计和计算技能（BBSRC核心技能-新的工作方式），以利用现有的大量和多样化的数据源的信息。这将有助于扩大和提高促进联合王国生物科学研究能力和生物经济所需的关键技能（生物科学研究中心核心技能-脆弱能力）。对抗AMR跨越多个BBSRC战略优先事项;该项目侧重于了解适用于AMR的分子变异的基础生物科学（战略研究优先事项3-健康生物科学），从而支持改善人类和动物健康（战略研究优先事项1-农业和粮食安全）。

Antimicrobial resistance (AMR) is a serious global health challenge, for both human and animal health. Antimicrobials are drugs that are used to fight pathogenic microbes (viruses, bacteria and parasites). The widespread use and misuse of these drugs has led to the emergence of resistance. One of the ways by which microbes become resistant to drugs and survive drug exposure is by undergoing genetic changes (mutations). There is an urgent need to identify and predict such mutations that lead to resistance before they become fixed in a given pathogen population. Technological advancements have led to rapid, cost effective, and extensive data collection in all aspects of health science. However, to fully exploit these information rich sources, it becomes imperative that corresponding computation techniques are also developed to enhance our understanding of complex biological problems. Artificial Intelligence through Machine learning (ML) offers a powerful way to expedite our understanding of these problems. As the name suggests, ML is a way to make computers learn from and discover patterns in data. The project aims to develop a robust and automated computational framework to identify new mutations in pathogens. Understanding the molecular consequences of these mutations will provide insight into the growing failure of antimicrobial therapy. Developing sophisticated machine learning approaches will accelerate the identification of new ways to detect and respond to AMR. This will have a direct impact in pathogen surveillance, aid stewardship and guide development of new interventions. This multi-disciplinary project integrates and enhances skills in maths, statistics and computation (BBSRC core skills - new ways of working) to leverage information from existing vast and diverse data sources. This will allow expansion and advancement of the vital skills necessary to promote the UK's bioscience research capabilities and bioeconomy (BBSRC core skills - vulnerable capabilities). Combatting AMR spans multiple BBSRC strategic priorities; the project focuses on understanding the basic bioscience of molecular variation (Strategic research priority 3 - bioscience for health) as applicable to AMR, thereby supporting improvements in both human and animal health (Strategic research priority 1 - agriculture and food security).