<![CDATA[Consumers of fermented products increasingly demand detailed information on product ingredients, quality, health benefits, and origin. Herein, we have chosen kombucha as a model for a fermented product. This study aims to establish the origin information of kombucha using clustering analysis of 16S metabarcoding datasets. We have downloaded and analysed datasets of kombucha 16S metabarcoding originating from 5 distinct places: Brazil, the United States, the United Kingdom, Turkey, and Thailand. We randomly selected datasets from the collection (n = 32) and analyzed them on the SHAMAN server to develop an initial microbiome profile. We implemented hierarchical agglomerative Clustering and found that Ward's method and the Chao distance produced the best cluster tree, which consistently separates kombucha into five distinct clades, reflecting their origin. We have extended our examination to include more datasets (n=13) to build the final cluster tree (total n = 45). We have also assessed the uncertainty of the final Clustering by pvclust in R. The pvclust cluster tree is comparable in topology to the final cluster tree built using Ward's method and Chao distance. The pvclust cluster tree features stable clades that are highly supported by AU (Approximately Unbiased) values (p-value ≥ 95%). Each kombucha was also placed correctly and consistently according to its respective origin. We have successfully conducted analyses and demonstrated that a simple clustering method, combining Ward's method and the Chao distance, is the most effective for classifying kombucha by origin using a 16S metabarcoding dataset.]]>
Copyrights © 2025
