The global emergence of novel pathogenic viruses presents an important challenge for research, as high biosafety levels are required to process samples. While inactivation of infectious agents facilitates the use of less stringent safety conditions, its effect on other biological entities of interest present in the sample is generally unknown. Here, we analyzed the effect of five inactivation methods (heat, ethanol, formaldehyde, psoralen, and TRIzol) on microbiome composition and diversity in samples collected from four different body sites (gut, nasal, oral, and skin) and compared them against untreated samples from the same tissues. We performed 16S rRNA gene sequencing and estimated abundance and diversity of bacterial taxa present in all samples. Nasal and skin samples were the most affected by inactivation, with ethanol and TRIzol inducing the largest changes in composition, and heat, formaldehyde, TRIzol, and psoralen inducing the largest changes in diversity. Oral and stool microbiomes were more robust to inactivation, with no significant changes in diversity and only moderate changes in composition. Firmicutes was the taxonomic group least affected by inactivation, while Bacteroidetes had a notable enrichment in nasal samples and moderate enrichment in fecal and oral samples. Actinobacteria were more notably depleted in fecal and skin samples, and Proteobacteria exhibited a more variable behavior depending on sample type and inactivation method. Overall, our results demonstrate that inactivation methods can alter the microbiome in a tissue-specific manner and that careful consideration should be given to the choice of method based on the sample type under study. IMPORTANCE Understanding how viral infections impact and are modulated by the microbiome is an important problem in basic research but is also of high clinical relevance under the current pandemic. To facilitate the study of interactions between microbial communities and pathogenic viruses under safe conditions, the infectious agent is generally inactivated prior to processing samples. The effect of this inactivation process in the microbiome is, however, unknown. Further, it is unclear whether biases introduced by inactivation methods are dependent on the sample type under study. Estimating the magnitude and nature of the changes induced by different methods in samples collected from various body sites thus provides important information for current and future studies that require inactivation of pathogenic agents.
Keywords: 16S RNA; DNA sequencing; human microbiome; viral inactivation.