Predicting Dietary Impact on Multiple Sclerosis-Related Symptoms With the Gut Microbiome: A Pilot Study Using Unsupervised Machine Learning

利用肠道微生物群预测饮食对多发性硬化症相关症状的影响:一项基于无监督机器学习的初步研究

阅读:1

Abstract

BACKGROUND: Multiple sclerosis (MS) is a neurodegenerative disease where dietary intervention has emerged as a potential adjunct treatment. Recently, the modified Paleolithic elimination (MPE) diet, also known as the Wahls diet, and the low-saturated fat (LSF) diet, also known as the Swank diet, were linked to reduced fatigue and improved quality of life (QoL) in the WAVES study (NCT02914964). However, how diet impacts these outcomes remains unclear. As diet impacts gut microbiota, we investigated whether the baseline gut microbiota can predict response to diet in people with MS (pwMS). METHODS: We performed fecal 16s rRNA sequencing to profile the microbiome changes associated with pwMS receiving the MPE (n = 11) and LSF diet (n = 12). Next, we utilized topic modeling, a machine learning technique, to determine whether baseline microbiome features predicted diet response in the combined MPE + LSF dietary cohort (n = 23). RESULTS: Specific genera significantly differed over time on both diets. On the MPE diet, Hungateiclostridiaceae, Ruminiclostridium, and Shuttleworthia decreased, while Coriobacteriaceae Collinsella decreased on LSF. Predictive machine-learning analysis associated a baseline microbiome enriched with Akkermansia, Bacteroides, and Barnesiella with fatigue response in the combined MPE + LSF cohort. For a non-response in Mental QoL improvement in the combined MPE + LSF cohort, our analysis associated an enrichment of Faecalibacterium and Alistipes at the start of the diet. DISCUSSION: Utilizing topic modeling, this pilot study identified baseline microbiota communities linked to improvements in fatigue and Mental QoL in pwMS on dietary intervention. These findings highlight the microbiota's role in dietary response and the potential for personalized nutrition. Given the small cohort and exploratory design, the results are hypothesis-generating and require validation in larger mechanistic studies.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。