EVALUATION OF TAXONOMIC CLASSIFICATION DATABASE FOR MICROBIOME ANALYSIS OF NANOPORE FULL-LENGTH 16S rRNA GENES

16S ribosomal RNA (rRNA) gene is a standard marker for bacterial identification. Recently, targeted amplicon sequencing of 16S hypervariable regions is a common technique for studying microbial communities. Bioinformatics tools for analyzing short-read sequences (e.g. QIIME2 and Mothur) have been developed and widely conducted. However, the microbial taxonomic classification based on one or two consecutive hypervariable regions is reliable only at a taxonomic genus level. The analysis of full-length 16S rRNA genes provides a higher resolution of microbial taxonomic classification at a species level, which is now possible by utilizing long-read nanopore sequencing technology. Nevertheless, the use of different 16S rRNA databases for taxonomic classification could produce inconsistent results. Here, we aim to compare and evaluate three different 16S rRNA databases for taxonomic classification of nanopore full-length 16S rRNA genes. The investigated 16S rRNA databases include i) NCBI-nucleotide, ii) NCBI-16S rRNA and iii) Silva138 databases. By conducting a mock microbial community dataset, curated taxa annotation of Silva138 resulted in the highest taxonomic classification accuracy, followed by the NCBI-nucleotide database and NCBI-16S rRNA database, respectively. A high amount of unclassified and uncultured microbes were revealed based on NCBI databases. The curated database of Silva provided a more comprehensive annotation of microbes at all taxonomic levels. We discuss and suggest a bioinformatics approach and database for analyzing microbiome data of nanopore full-length 16S rRNA sequencing for high-resolution microbial taxonomic classification at a species level. The selection of a taxonomic database provides a more informative annotation of the identified microbes.