Gene copy normalization of the 16S rRNA gene cannot outweigh the methodological biases of sequencing

Abstract

The 16S rRNA gene is the golden standard target of sequencing to uncover the composition of bacterial communities but the presence of multiple copies of the gene makes gene copy normalization (GCN) inevitable. Even though GCN resulted in abundances closer to the metagenome, it should be validated by communities with known composition as both amplicon and shotgun sequencing are prone to methodological biases. Here we compared the composition of three mock communities to the composition derived from 16S sequencing without and with GCN. In all of them, the 16S composition was different from the mock community and GCN improved the picture only in the community with the lowest Shannon diversity. Albeit with low abundance, half of the identified genera were not present in the mock communities. Our approach provides empirical evidence to the methodological biases introduced by sequencing that was only counteracted by GCN in the case of low α-diversity, potentially due to the small number of bacterial taxa with known gene copy numbers. We thus cannot recommend the use of GCN moving forward and it is questionable whether a complete catalogue of 16S rRNA copy numbers can outweigh the methodological biases of sequencing.