Effects of the long-term storage of human fecal microbiota samples collected in RNAlater.
Tap J, Cools-Portier S, Pavan S, Druesne A, Öhman L, Törnblom H, Simren M, Derrien M
Sci Rep. Jan 2019. doi: 10.1038/s41598-018-36953-5
COMMENT: 16S rRNA gene sequencing of fecal samples is an effective and non-invasive method to study the gut microbiome. An appropiate storage of sample material is essential to avoid the introduction of bias in microbial community composition. In this work, the authors evaluate the influence of long-term storage on the microbial composition of human stool samples using RNAlater as preservative.
The aim of this study was to evaluate the effects on the human fecal microbiota of a long-term storage (5 years) of fecal samples collected in RNA later and to compare these effects with those of technical and biological variability (intra- and inter-subject variability).
Changes in alpha-diversity (...) between follow-up and reference samples, were larger within and between subjects (intra- and inter-subject respectively), than between sequencing runs and before and after long-term storage. Furthermore, similar results were obtained for changes in beta-diversity (...) at the genus and OTU levels . The inter-sequencing run effect was not signicantly different from the fecal storage effect. Inter-sequencing run and fecal storage effects were signicantly smaller than the intra-subject variability.
Higher initial richness tended to be associated with a moderately higher sensitivity to change during long-term storage. Bacterial counts were positively correlated with species richness, but not associated with fecal sample instability after long-term storage. Some changes in sample stability were observed during long-term storage, particularly for low prevalent and low abundant members.
The long-term storage of stool samples at −80 °C had little effect on the integrity of the gut microbiota, although minor changes in the relative abundance of some species correlated with initial microbiota composition were observed. However, these changes were always smaller than inter-individual and inter-sequencing run variability.