Fighting Infection and AMR in broiler farming: AI, omics and smart sensing for diagnostics, treatment selection and gut microbiome improvement

与肠道感染作斗争，同时遏制抗菌素耐药性的上升，是当代肉鸡养殖的主要挑战之一，对鸟类和消费者的健康都有影响。未来更好的监测、诊断和治疗选择解决方案的关键是更好地了解鸟类的肠道微生物群，探索其共生菌和机会性病原体种群因感染、治疗和抗性性状的发展而发生的变化。在这个项目中，我们计划探索肉鸡肠道微生物群，重点研究与鸟类胃肠道中常见的病原体有关的感染和耐药性：产气荚膜梭菌、盲肠球菌、大肠杆菌和沙门氏菌，我们也涵盖了与病毒共同感染导致肠道微生物群失调的情况。我们考虑对8类抗生素的耐药性/易感性：四环素类、磺胺类、β -内酰胺类、氟喹诺酮类、多粘菌素类、大环内酯类、二氨基嘧啶类、氨基糖苷类，这些抗生素作为治疗药物在英国广泛使用。我们计划从农场、饲料和鸟类中收集大量异构数据，涵盖正常生产时期和感染事件。数据将包括微生物分析结果、全基因组测序、鸟枪宏基因组学和对粪便样本进行的表型分析、农场管理实践，以及环境传感器数据和鸟类成像。我们建议使用机器学习和云计算来执行大规模数据挖掘，并最终解开可观察变量之间可能相互作用的网络，跟踪肉鸡的生命周期，并捕捉感染，治疗和单药或多药耐药性的发展。获得的知识可能为选择可观察变量提供提示，这些变量可作为生物标志物，即在未来的牲畜实时监测解决方案中可作为目标，以检测/预测感染或抗性性状的存在/叛乱，并支持精确诊断和定制治疗选择。研究结果还可能提出改善鸟类肠道微生物群的途径，例如通过饲料添加剂，使其对感染更强大，同时抑制耐药性的发展。

The fight against enteric infections while containing the uprise of antimicrobial resistance, represents one of the major challenges in contemporary broiler farming, with repercussions on both bird and consumer's health. Key to future, better solutions for surveillance, diagnostics and treatment selection, is to gain an improved understanding of the bird's gut microbiome, exploring the modifications its population of commensals and opportunistic pathogens undergo as a consequence of infection, treatment and development of resistant traits. In this project, we plan to explore the broiler gut microbiome, focusing on infection and resistance in relation to pathogens typically found in the gastrointestinal tract of the birds: Clostridium perfringens, Enterococcus cecorum, Escherichia coli and Salmonella spp. We cover also scenarios of co-infection with viruses causing dysbiosis of gut microbiome. We consider resistance/susceptibility to 8 classes of antibiotics: tetracyclines, sulphonamides, beta-lactams, fluoroquinolones, polymyxins, macrolides, diaminopyrimidines, aminoglycosides, whose use as therapeutics is diffused in the UK. We plan to collect a large amount of heterogeneous data from farms, feed and birds, covering normal production periods and infection events. Data will include results of microbiological analysis, whole-genome sequencing, shotgun metagenomics and phenotyping performed on faecal samples, on-farm management practices, as well as environmental sensor data and bird imaging. We propose to use machine learning and cloud computing to perform large-scale data mining and ultimately unravel the network of possible interactions amongst the observable variables, following broilers along their life cycle, and capturing episodes of infection, treatment and development of single or multi-drug resistance. Acquired knowledge may provide hints at the selection of observable variables acting as biomarkers, i.e, targetable by future solutions for real-time livestock monitoring, to detect/forecast infection or the presence/insurgence of resistant traits, and to support precision diagnostics and bespoke treatment selection. The results may also suggest routes to improve the birds gut microbiome, for example via feed additives, making it more robust to infection while at the same time inhibiting the development of resistance.