The Signal Through the Noise: Identifying Subsurface Associated Microorganisms from Oilfield Samples

In the age of big data, discerning signal from noise has become paramount. With the rapid acquisition of DNA sequence data being utilized to interpret environments, faith is placed in the fact that such data is representative of the intended environment. Collection of environmental microbial data follows two routes: 1) direct collection: scooping up soil, or filtering water onsite and extracting DNA, or 2) in the case of oil reservoir samples, a more indirect collection, sampling production wells, separators, or holding tanks. This indirect method introduces additional microbial communities, by the time the researcher accesses a sample, many microbial community signals are present. Additionally, the difficulty in collection from these production sites leads to the potential for small sample set size. Interpretations of the microbial community structure, function, and community interactions can be clouded by the contaminating communities, and inappropriate emphasis on outliers can be leveraged to promote false claims. We propose a method by which, with appropriate sample set size, we can begin to separate out the true signal of the environment of interest from the contaminating microbial "noise". Production wells and topside separators were sampled from multiple reservoirs and the microbial communities were analyzed through Illumina sequencing of the 16S rRNA gene. We then constructed a co-occurrence network, removed weak correlations, identified groups of highly-interactive organisms and assigned value to each. Groups of organisms were then related back to the environmental parameters for each sample. Using this technique we were able to discern that a large proportion of our "oil reservoir" community was not, in fact, from the subsurface. Additionally, we were able to assign species of unknown origin to an environment. Not taking microbial community samples at face value when working with these types of samples is integral in elucidating microbial communities from hard to reach environments.