Abstract
Abstract Background The importance of microbial community profiling has been increasingly recognized in biological and environmental research. While metabarcoding has been widely used for such analysis by targeting specific DNA sequences as markers, shotgun sequencing has been proposed as an alternative method because the analysis of whole genomes can potentially reduce biases introduced by targeted approaches. However, it is largely unknown whether shotgun sequencing may provide improved precision for qualitative taxonomic identification and quantitative abundance estimation compared with metabarcoding with housekeeping gene markers, such as the rpoB gene. Furthermore, the comparative performance of various bioinformatics pipelines for shotgun data analysis remains uncertain. In this study, we evaluated the performance of rpoB metabarcoding and shotgun sequencing coupled to various bioinformatic pipelines to describe the bacterial diversity of artificially generated mock bacterial communities, which included eukaryote gDNA intentional contamination or whole-genome amplification. For shotgun sequencing, the Assembly-Binning-Method and k-mer-based approaches were evaluated. Results For taxonomic profiling, the Assembly-Binning-Method and rpoB metabarcoding exhibited comparable sensitivity and precision, whereas k-mer approaches produced a notably high number of false negatives. In some cases, the Assembly-Binning-Method improved taxonomic resolution compared with rpoB metabarcoding by identifying taxa at the species level rather than the genus level. Regarding the quantification of microbial composition, the Assembly-Binning-Method consistently showed a higher correlation with expected values and a lower dissimilarity index than rpoB metabarcoding. The use of three sets of reference genomes to calculate depth coverage did not systematically affect the precision of the Assembly-Binning-Method. Conclusions These results demonstrate that although shotgun sequencing and rpoB metabarcoding have nearly equivalent accuracy in taxonomic profiling, shotgun sequencing has better taxonomic resolution and outperforms rpoB metabarcoding in quantitative estimation of microbial community abundance using the Assembly-Binning approach.