16S rRNA gene amplicon sequencing on Miseq using mock communities of bacteria and archaea

Target gene amplicon sequencing, especially on Miseq, is prevalent in microbial ecology researches. However, the artificial OTUs, quantitation and bias issues are concerns in application of this method in more widen fields. Using mock communities of full length 16S rDNA of 33 bacterial strains and 24 archaea strains, a systematic analysis was done to identify the sources of artificial OTUs, strain quantitation variation and biases of this method. Both bacteria and archaea strains were divided into three groups based on their GC contents, high, moderate, and low. Three mock communities were built from both the bacteria and archaea strains: Mock B1, B2, and B3, and Mock A1, A2, and A3. Mock B1 and A1 were with all strains of the same abundances, Mock B2 and A2 were with the high GC content strains as the high abundance group, the moderate GC content strains as the middle abundance groups, and the low GC content strains as the low abundance groups. Mock B3 and A3 were with the high GC content strains as the low abundance group, the moderate GC content strains as the middle abundance groups, and the low GC content strains as the high abundance groups. PCR libraries were generated with non-phasing method, one-step-phasing method, and two-step-phasing method. PCR library generation were replicated 24 time for each mock community and each method. For phasing methods, each of the 8 spacer groups (spacer 0-7) were replicated three times. Then, the libraries were sequenced with Illumina Miseq.