Shotgun sequencing and metabarcoding for bacterial profiling and quantitication

Accurate profiling of microbial communities is crucial for biological and environmental research. While metabarcoding, such as rpoB gene-targeted sequencing, is widely used for this purpose, shotgun sequencing, which analyzes whole genomes, offers a promising alternative by potentially reducing biases inherent in targeted approaches. Previous studies comparing shotgun sequencing with 16S rRNA metabarcoding have shown better species-level resolution with shotgun sequencing. However, a systematic comparison of rpoB metabarcoding and shotgun sequencing, particularly for quantification, remains unexplored. In this study, we compared rpoB metabarcoding and shotgun sequencing using mock bacterial communities, employing various bioinformatics pipelines, including the Assembly-Binning-Method and k-mer-based approaches. For taxonomic profiling, the Assembly-Binning-Method and rpoB metabarcoding displayed similar sensitivity and precision, while k-mer approaches produced a high number of false negatives. The Assembly-Binning-Method sometimes improved taxonomic resolution by identifying taxa at the species level instead of genus. In terms of microbial composition quantification, the Assembly-Binning-Method showed higher correlation with theoretical values and lower dissimilarity compared to rpoB metabarcoding. The choice of reference database and the calculation of depth coverage using different reference genomes did not significantly affect the precision of the Assembly-Binning-Method