The Source and Evolutionary History of a Microbial Contaminant Identified Through Soil Metagenomic Analysis

Abstract

In this study, strain-resolved metagenomics was used to solve a mystery. A 6.4-Mbp complete closed genome was recovered from a soil metagenome and found to be astonishingly similar to that of Delftia acidovorans SPH-1, which was isolated in Germany a decade ago. It was suspected that this organism was not native to the soil sample because it lacked the diversity that is characteristic of other soil organisms; this suspicion was confirmed when PCR testing failed to detect the bacterium in the original soil samples. D. acidovorans was also identified in 16 previously published metagenomes from multiple environments, but detailed-scale single nucleotide polymorphism analysis grouped these into five distinct clades. All of the strains indicated as contaminants fell into one clade. Fragment length anomalies were identified in paired reads mapping to the contaminant clade genotypes only. This finding was used to establish that the DNA was present in specific size selection reagents used during sequencing. Ultimately, the source of the contaminant was identified as bacterial biofilms growing in tubing. On the basis of direct measurement of the rate of fixation of mutations across the period of time in which contamination was occurring, we estimated the time of separation of the contaminant strain from the genomically sequenced ancestral population within a factor of 2. This research serves as a case study of high-resolution microbial forensics and strain tracking accomplished through metagenomics-based comparative genomics. The specific case reported here is unusual in that the study was conducted in the background of a soil metagenome and the conclusions were confirmed by independent methods.

Genome-resolved metagenomics of eukaryotic populations during early colonization of premature infants and in hospital rooms

“…This study made use of many different previously analyzed infant datasets. These datasets have previously published descriptions of the study design, patient selection, and sample collection, and are referred to as NIH1 [51, 52], NIH2 [19], NIH3 [53], NIH4 [54], Sloan2 [19], and SP_CRL [55]. Infants were chosen for inclusion in this study irrespective of fungal disease state.…”

“…This study also involved the collection and processing of an additional 269 samples from 53 infants. Newly collected infant fecal samples followed the same sample collection and DNA extraction protocol as described previously [53, 56]. Metagenomic sequencing of newly collected infant fecal samples was performed in collaboration with the Functional Genomics and Vincent J. Coates Genomics Sequencing Laboratories at the University of California, Berkeley.…”

Identification and Removal of Contaminant Sequences From Ribosomal Gene Databases: Lessons From the Census of Deep Life

“…However, through the use of metagenomics ( Brazelton et al, 2012 ), researchers have shown that members of the Betaproteobacteria are in fact ubiquitous and ecologically important members of alkaline (pH > 10) serpentinite environments, which was subsequently verified by culture-dependent studies ( Suzuki et al, 2014 ). Additionally, Olm et al (2017) recovered Delftia genomes from soil metagenome assemblies and were able to show these genomes were from contamination using comparative genomics. Thus, the ability to trace contamination with high precision using techniques that identify strain level variation, gene additions or loss, and the presence of plasmids is at the forefront of this detective work.…”

Strain-level overlap between infant and hospital fungal microbiomes revealed throughde novoassembly of eukaryotic genomes from metagenomes

“…Further, the samples from the same infant were all processed at the same time using the same reagents, yet abundances change substantially across the sample series. Additionally, the two Purpureocillium genomes reconstructed from gut samples from different infants are distinct ( Figure 4 ); one would expect the sequences to be more similar if derived from the same reagent contaminants (Olm et al, 2017b) .…”