Purpose: The purpose of this study was to investigate NGS as a reliable method for qualitative community analysis and to characterize interlab variability.
Methods: Three independent laboratories were contracted to conduct NGS-based bacterial community analysis of two samples (in triplicate). The first sample was prepared in-house and represented a known composition of Mycobacterium smegmatis (73.1%), Pseudomonas aeruginosa (24.3%), Escherichia coli (1.2%), Staphylococcus aureus (1.0%), and Bacillus subtilis (0.3%). The second sample consisted of an environmental swab taken from a kitchen floor which represented a sample of unknown composition. Labs were instructed to perform bacterial community analysis using their own methods and procedures. The results from each lab were compared for accuracy and consistency across replicates, as well as consistency across labs.
Results: For known samples, each lab failed to identify the presence of M. smegmatis. The remaining species were detected at varying levels, for example, B. subtilis was detected at levels ranging from 0.3-17.3%. For field samples, each lab was in agreement on the top four contributors to the community, with Psychrobacter spp. (16.5-48.5%) and Enterobacteriaceae (14.2-25.5%) being the most abundant organisms. Large variations arose after the top four organisms, for example, one lab identified Rothia spp. being present at 7.5% while other labs did not detect the organism.
Significance: NGS may provide insight into community membership, but there is considerable variability in the results from identical samples. The results of the study highlight interlab variability while demonstrating the need to harmonize sample preparation and analytical techniques used in NGS.