There are multiple software programs available for operational taxonomical unit (OTU) clustering and multiple 16S rRNA gene sequence databases, and the statistical properties of statistical approaches ultimately depend on which method is used. The statistical power should be maximized when the OTU clustering results are the most accurate, but the best strategy for OTU clustering remains unclear. Multiple studies have compared the accuracy of databases by using a mock community whose microbial composition is known, and the EzTaxon database was reported to be the most accurate among the existing databases, including the Silva and Greengenes databases. But EzTaxon database is not widely used and lacks in terms of the number of OTUs they can identify. Recently, amplicon sequence variant (ASV) is proposed for constructing metagenomics dataset and it has its obvious strengths in its consistent labeling and reference independence. In this study, we developed metrics for comparing datasets based on various clustering methods and databases. Then, we proposed a novel microbiome association test based on multiple databases and clustering methods. We observed that datasets based on different databases were similar when the share the same methods of clustering (Figure 1). Also, compared to other methods, ASV's dataset properties were unique.<br/>

Multiple clustering methods and databases-based microbiome association test

ACS Fall 2020 Virtual Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.