Over the past decade, next-generation sequencing of the human microbiome has yielded insights into a variety of diseases, but the results have often been conflicting or inconclusive.
The misinterpretation of data in the field might arise from artifacts introduced by next-generation sequencing protocols such as library preparation, which converts nucleic acid material into standard libraries suitable for loading onto a sequencing instrument.
To address this issue, researchers in the journal PNAS compared whole-genome sequencing data from complete microbial communities using four different library preparation protocols.
Two of these methods required PCR, which amplifies nucleic acid fragments in the library prior to sequencing. The authors compared the methods using stool samples from a healthy volunteer treated with antibiotics, as well as a synthetic mock community composed of a mixture 20 microbial genomic DNAs.
The two methods that did not require PCR resulted in lower error rates and higher-quality reads for the mock community, compared with the PCR-based methods. Moreover, the four different libraries showed significant variation in the relative abundance of microbial members of the mock community and the stool samples.
According to the authors, the findings highlight the need for consistency across next-generation sequencing protocols to improve the interpretation of microbiome data in studies of human health and disease.