CloudLab: Analyzing Complex Microbial Responses to Dynamic Environments

Access to clean, reliable water supplies is critical to our quality of life and our economy, yet across the country thousands of hazardous waste sites are so heavily contaminated that the underlying groundwater doesn’t meet drinking water standards. Measuring contamination in the environment is critical to human health and to understanding complex subsurface systems, but current testing is mostly limited to sporadic sample collection for laboratory analysis. Not only is this process costly, but it is inefficient, making it difficult to monitor groundwater with high spatial or temporal resolution. As a result, current methods often do not capture the full complexity of how contaminants behave in the environment. The main objective of this SBIR proposal is to develop the computational technology required to support a scalable and flexible sensor network that can continuously monitor many different contaminants. Given the priorities of the DOE Biological and Environmental Research (BER) and Advanced Scientific Computing Research (ASCR) programs, our goal for this proposal will be to develop algorithms and associated software that interface with our sensor’s complex imaging data to avoid computational and bandwidth bottlenecks while still providing end users the ability to efficiently validate, navigate, and analyze sensor results in real time. BER focus areas would benefit from spatial and temporal mapping of water quality. For example, the Subsurface Biogeochemistry Research Pro- gram states one of the knowledge gaps is to understand the “flux of contaminants and nutrients between groundwater and surface water.”A successful outcome of this Phase I proposal willbe algorithms and software that support a scalable biosensor network capable of real-time quantification of arsenic, cadmium, nitrogen, and phosphorus. A Phase II proposal would focus on real-world performance evaluations of a sensor network via deployment in areas of concern and tailoring our platform to the needs of the DOE and other potential end users. The ability to detect many targets simultaneously and to do so in real time is a major advantage of our sensor over existing technologies. While there are several online sensors on the market, they are typically characterized by large capital investments (around $50k), high maintenance costs, and cumbersome physical footprints, and they typically can only sense one or two targets. Our biosensor will be able to quickly and accurately measure the concentration of many different contaminants in water continuously for over a month without requiring subjective interpretation, extensive calibration, or regular oversight. A simple monthly cartridge and media swap will enable the sensor to run indefinitely, reporting the contaminant concentrations either on the unit itself or transmitted remotely. In addition, it will be small and cheap enough to deploy many units to enable mapping of real-time contamination throughout an area of concern, which will contribute to predictive understanding of how contaminants move through a watershed. Developing software for our advanced toxin sensor to address the limitations presented by current technologies would allow monitoring to become more affordable, continuous, and field- deployable. The CloudLab software environment proposed here will be highly marketable to a wide range of end-users, as it will meet the increasing need for data to inform management decisions aimed at reducing contaminant concentrations in the environment, tracking the progression of contamination plumes, and targeting investments in remediation efforts at legacy waste sites.