查看文章

Sequencing of the 16S rRNA gene by Illumina next-generation sequencing is broadly used in microbiome studies. Different hypervariable regions of the 16S rRNA gene, V3V4 (amplified with primers 341F–805R) or V4 (V4O; primers 515F–806R), are selected, depending on the targeted resolution. However, in population-based clinical studies, combining V3V4 and V4 data from different studies for a meta-analysis is challenging. Reads generated by short-read (150-bp) high-throughput sequencing platforms do not fully recover the V4 region read-length. Here, we evaluated the compatibility of 16S rRNA V3V4 and V4 amplicons for microbiome profiling. We compared taxonomic compositions obtained by the analysis of V3V4 and V4 amplicons, and V4 fragments trimmed from V3V4 amplicons. We also evaluated an alternative V4 region (V4N; primers 519F–798R) designed for efficient stitching with 150-bp paired-end sequencing. First, we simulated a global investigation of environmental prokaryotes in silico. This revealed that V4O primers recovered the highest proportion of fragments (81.7%) and most phyla, including archaea. Empirical sequencing of standard (mock) and human fecal samples revealed biased patterns of each primer that were similar to the ones determined by in silico simulation. Further, for human fecal microbiome profiling, the between-sample variance was greater than the systematic bias of each primer. The use of trimmed V4 fragments and single-end amplicons resulted in the same systematic bias. In conclusion, paired-end V4O sequencing yielded the most accurate data for both, simulation and mock community …