Optimized identification of therapeutic bacterial strains in ulcerative colitis

项目总结 溃疡性结肠炎(UC)是一种慢性肠道炎症，被认为是由遗传因素引起的 以及环境因素。治疗方案在诱导和维持方面只有部分效果 缓解，目前还没有治愈UC的方法。最近的研究发现，UC患者的微生物群是 与健康对照组相比截然不同，这表明微生物组可能是一种有前途的治疗方法 目标。粪便微生物区系移植(FMT)在治疗由以下原因引起的结肠炎方面取得了极大的成功 艰难梭菌感染，进一步证明了细菌疗法的潜力。最近的 完成的FOCUS(粪便微生物区系移植治疗溃疡性结肠炎)试验表明，FMT 27%的UC患者可以从健康供体中获得临床缓解，而服用安慰剂的患者只有8% 一群人。然而，我们仍然缺乏对FMT能够调节的确切机制的了解 疾病发病机制。这项建议的首要假设是FMT后UC缓解 由特定细菌诱导，而准确鉴定细菌菌株将是至关重要的 开发治疗UC的微生物疗法。为了验证这一假设，我们将首先开发一个新的合奏 方法从深层元基因组数据中鉴定细菌种类和菌株。这种方法结合了软件 工具使用投票算法，该算法根据每个工具的精度进行加权。这种方法的准确性 将使用UC患者的微生物培养收集、计算机模拟和细菌模拟进行测试 以社区为黄金标准。然后，我们将把这种方法应用于从 重点研究以确定与UC缓解相关的细菌菌株。这将被执行 通过使用已建立的监督学习技术，以及通过一种基于联合学习的新方法， 识别协同作用的细菌集群的发生网络分析。根据这些结果， 我们最终将使用结肠炎的诺维菌学模型来验证我们的发现。预防性和/或治疗性 候选细菌的潜力将通过将它们接种到定居在那里的灵知菌小鼠来测试 通过转移幼稚T细胞诱导的UC患者和结肠炎患者的微生物区系。我们的建议将是 首先在一项随机试验中解决特定细菌如何调节疾病进展的问题 UC中的FMT通过开发更灵敏和准确的菌株表征方法 可以通过实验进行验证。我们鉴定治疗性细菌的策略的合理设计将产生 高度翻译的结果，可用于指导未来在UC的临床试验。

PROJECT SUMMARY Ulcerative colitis (UC) is a chronic gut inflammatory condition thought to be caused by a combination of genetic and environmental factors. Therapeutic options are only partially effective in inducing and maintaining remission, and there is currently no cure for UC. Recent studies have found the microbiome of UC patients is distinct compared to that of healthy controls, suggesting that the microbiome could be a promising therapeutic target. Fecal microbiota transplant (FMT) has been extremely successful in the treatment of colitis caused by Clostridium difficile infection, further demonstrating the potential for bacterial therapeutics. The recently completed FOCUS (Faecal Microbiota Transplantation in Ulcerative Colitis) trial has demonstrated that FMT from healthy donors can induce clinical remission in 27% of UC patients compared to only 8% in the placebo group. However, we still lack an understanding of the exact mechanisms by which FMT is able to modulate disease pathogenesis. The overarching hypothesis of this proposal is that UC remission after FMT is induced by specific bacteria, and that accurate identification of bacterial strains will be paramount to develop microbial therapeutics for UC. To test this hypothesis, we will first develop a novel ensemble method to identify bacterial species and strains from deep metagenomic data. This method combines software tools using a voting algorithm that is weighted based on the accuracy of each tool. The accuracy of this method will be tested using microbial culture collections from UC patients, in silico simulations, and a bacterial mock community as gold standards. We will then apply this approach to metagenomic data generated from the FOCUS study in order to identify bacterial strains that are associated with UC remission. This will be performed by using established supervised learning techniques, as well as through a novel method based on co- occurrence network analysis that identifies clusters of bacteria that act synergistically. Based on these results, we will finally validate our findings using a gnotobiotic model of colitis. The prophylactic and/or therapeutic potential of candidate bacteria will be tested by inoculating them to gnotobiotic mice colonized with the microbiota of UC patients and with colitis induced through the transfer of naïve T cells. Our proposal will be the first to address the question of how specific bacteria modulate disease progression in a randomized trial of FMT in UC through the development of more sensitive and accurate methods for strain characterization that can be experimentally validated. The rational design of our strategy to identify therapeutic bacteria will produce highly translational results that can be used to guide future clinical trials in UC.