Evaluating the Efficiency of 16S-ITS-23S operon Sequencing: A Comparison of Primer Pairs, Sequencing Platforms, and Taxonomic Classifiers

Background The field of 16S rRNA-targeted metagenetics has been enhanced through the improved accuracy of long-read sequencing. More specifically, recent advances have facilitated the transition from short-read sequencing of 16S rRNA gene regions to full-length sequencing of the entire 16S gene (~1500 bp) and, in turn, sequencing of the 16S, Internal Transcribed Spacer (ITS), and 23S regions covering a DNA region known as the ribosomal RNA operon (RRN) (~4500 bp). These technological advances offer the potential to achieve at least species-level resolution when analysing microbiomes, increasing interest in RRN sequencing. However, before widespread adoption of this approach can occur successfully, a thorough assessment of its strengths and limitations is necessary. Results This study assesses the effects of RRN primer pairs and sequencing platforms on RRN sequencing, while also aiming to benchmark taxonomic classification methods. In this context, we study the effect four RRN primer combinations; four mock communities, three sequencing platforms (PacBio, Oxford Nanopore Technologies, and Illumina), two classification approaches (Minimap2 alignment and OTU clustering), and four RRN reference databases (MIrROR, rrnDB, and two iterations of FANGORN) alongside two 16S databases (Greengenes2 and SILVA). Our study reveals that choice of primer pair and sequencing platform do not substantially bias the taxonomic profiles provided by RRN sequencing for a majority of the mock communities. However, community composition was identified as a confounding factor. The classification method significantly impacts the accuracy of species-level taxonomic assignment. Applying Minimap2 in combination with the FANGORN database was found to provide the most accurate profile for most microbial communities, irrespective of sequencing platform. Conclusions Long-read sequencing of the RRN operon provides species-level resolution surpassing that of Illumina-based 16S rRNA gene sequencing. Our findings advocate for the use of RRN sequencing in species-level microbial profiling. We extensively benchmark the factors involved to provide a valuable resource, aiding the advancement and adoption of RRN sequencing, while highlighting some ongoing challenges.