16S rRNA gene sequencing and machine learning reveal correlation between drug abuse and human host gut microbiota

Over the past few years, there has been increasing evidence highlighting the strong connection between gut microbiota and overall well‐being of the host. This has led to a renewed emphasis on studying and addressing substance use disorder from the perspective of brain‐gut axis. Previous studies have suggested that alcohol, food, and cigarette addictions are strongly linked to gut microbiota and faecal microbiota transplantation or the use of probiotics achieved significant efficacy. Unfortunately, little is known about the relationship between drug abuse and gut microbiota. This paper aims to reveal the potential correlation between gut microbiota and drug abuse and to develop an accurate identification model for drug‐related faeces samples by machine learning. Faecal samples were collected from 476 participants from three regions in China (Shanghai, Yunnan, and Shandong). Their gut microbiota information was obtained using 16S rRNA gene sequencing, and a substance use disorder identification model was developed by machine learning. Analysis revealed a lower diversity and a more homogeneous gut microbiota community structure among participants with substance use disorder. Bacteroides, Prevotella_9, Faecalibacterium, and Blautia were identified as important biomarkers associated with substance use disorder. The function prediction analysis revealed that the citrate and reductive citrate cycles were significantly upregulated in the substance use disorder group, while the shikimate pathway was downregulated. In addition, the machine learning model could distinguish faecal samples between substance users and nonsubstance users with an AUC = 0.9, indicating its potential use in predicting and screening individuals with substance use disorder within the community in the future.

在过去几年中，越来越多的证据凸显了肠道微生物群与宿主整体健康之间的紧密联系。这使得从脑 - 肠轴的角度研究和解决物质使用障碍受到了新的重视。先前的研究表明，酒精、食物和香烟成瘾与肠道微生物群密切相关，粪便微生物群移植或使用益生菌取得了显著效果。不幸的是，人们对药物滥用与肠道微生物群之间的关系知之甚少。本文旨在揭示肠道微生物群与药物滥用之间的潜在相关性，并通过机器学习为与药物相关的粪便样本建立准确的识别模型。从中国三个地区（上海、云南和山东）的476名参与者收集了粪便样本。利用16S rRNA基因测序获取了他们的肠道微生物群信息，并通过机器学习建立了一个物质使用障碍识别模型。分析显示，物质使用障碍参与者的肠道微生物群多样性较低，群落结构更均匀。拟杆菌属、普氏菌属_9、粪杆菌属和布劳特氏菌属被确定为与物质使用障碍相关的重要生物标志物。功能预测分析表明，柠檬酸循环和还原柠檬酸循环在物质使用障碍组中显著上调，而莽草酸途径则下调。此外，该机器学习模型能够区分物质使用者和非物质使用者的粪便样本，曲线下面积（AUC）= 0.9，这表明它在未来预测和筛选社区内物质使用障碍个体方面具有潜在用途。