In addition to nucleic and mitochondrial sources of cell-free DNA (cfDNA) found in human blood, accumulating evidence indicates that commensal microbiome-derived DNA may also contribute to the composition of cfDNA. However, much of this evidence is questionable due to limited use of rigorous experimental controls that account for contamination that invariably affect low-biomass microbiome studies.
Here, we applied a 16S-rRNA-gene based approach to determine if the detection of bacterial cfDNA from plasma was possible and could be discriminated from background contamination from 80 metastatic melanoma patients showing no signs of infection. A 16S-rRNA-gene droplet digital PCR (ddPCR) assay showed that the concentration of bacterial cfDNA is low (mean of 1,312 16S rRNA gene copies/ml of plasma) yet significantly above the levels detected from DNA Extraction Negative Controls (DENC; mean of 877 16S rRNA gene copies/ml of plasma; t-test p-value=0.017) where nuclease-free water was the input used for DNA extraction. To compare the microbiome composition of plasma and DENCs we then performed 16S rRNA gene sequencing. The microbial community of plasma is heavily affected by batch effects most likely due to laboratory and kit contaminants introduced during DNA extraction. Supporting this observation, beta-diversity analysis of a subset of patients that had matching plasma, stool and saliva samples, showed that the community structure of plasma overlapped with its corresponding DENCs, while the community structure of stool and saliva did not.
Through the application of a series of informatic-based decontamination criteria including the filtering of Operational Taxonomic Units (OTUs) with batch dependant abundances (using limma R package) or those with a higher prevalence in DENC vs plasma samples across batches (using decontam R package), a small number of “candidate genuine plasma OTUs” were identified and are currently being validated via ddPCR. These preliminary results show that despite the inadequacy of current cfDNA isolation standards in dealing with the challenges associated with a low-biomass microbiome, a genuine circulating bacterial cfDNA signal can be recovered.