Comprehensive evaluation of sequencing variation caused by sampling site, sequencer and library preparation method in microbiome research

The gut microbiome sequencing has made highly diverse microbiota profiling possible and push the frontiers of Omics studies in human biology. However, there are thorny issues in the microbiome sequencing quality control aspect. How the sampling site affects sequencing results and the extent of variations between sequencers with respect to whole-metagenome shotgun (WMS) library preparation protocols remains a neglected area of research. It is also unknown whether sequencing results from WMS protocols are comparable with those from the 16s amplicon protocol.In the current study, we compared biological replicates from different sampling sites, evaluated three widely adopted sequencers (HiSeq2500, X10 and NovSeq) with the WMS protocol, and further compared the 16s data with the WMS library protocol. Finally, several algorithms were employed and compared to improve the population-wide comparability between the 16s and WMS protocols. We found heterogeneity between two sampling sites, comparable sequencing results generated by three sequencers, and major taxonomic profiling differences between the 16s protocol and WMS protocol. Additionally, our approach provides an effective tool to enhance the population-wide comparability between 16s and WMS data. Our orthogonal study not only the irreproducibility behind sampling sites, sequencer, and library preparation protocol, but also provided an effective way to reduce the biased readout of the datasets. We hope this work future can constitute a fundamental reference for WMS sequencer choice, sample preparation, and integrated analysis across 16s and WMS datasets.