About This Project

Hypersaline lakes contain valuable pigment-producing microbes, but mapping them requires extensive sampling. We ask whether drone-based hyperspectral imaging can identify these communities from their unique light signatures. We hypothesize that pigment signatures will predict microbial composition when validated with DNA sequencing and environmental data. This study could enable rapid, noninvasive mapping of Great Salt Lake microbial diversity.

Ask the Scientists

What is the context of this research?

Hypersaline lakes host microorganisms called extremophiles that thrive in conditions such as intense sunlight, low oxygen, and extreme salinity that are lethal to most life. Many produce colorful pigments that protect them from sunlight or help them convert light into energy. These include bacterioruberin and bacteriorhodopsin from salt-loving microbes called haloarchaea, along with other potentially valuable pigments produced by Dunaliella algae and Salinibacter bacteria (Rodrigo-Baños et al., 2015). Because these pigments interact with light in distinct ways, they may act as optical “fingerprints” detectable by imaging instruments (Wurtsbaugh et al., 2017). We hypothesize that visible and near-infrared (VNIR) imaging, which measures reflected light beyond what the human eye can see, can be combined with DNA sequencing and environmental measurements to predict which pigment-producing microbes are present and enable large-scale monitoring of the lake.

What is the significance of this project?

This project advances four edges. Scientifically, it establishes drone-based hyperspectral imaging as a quantitative method for resolving carotenoid and retinal chromophore distributions in microbial communities, bridging the gap between destructive sampling and coarse satellite remote sensing. Ecologically, it produces a spatially-resolved baseline of Great Salt Lake halophile community structure at a critical moment — the lake has lost roughly two-thirds of its surface area since 1850, threatening these communities before they have been comprehensively characterized. Biotechnologically, the ML classifier framework enables rapid identification of microhabitats enriched in commercially relevant carotenoids and retinal proteins used in antioxidant, cosmetic, and optogenetic applications. Methodologically, this transfers directly to other extremophile systems, providing infrastructure for spatial microbial biogeography.

What are the goals of the project?

This project will deploy a calibrated drone-mounted visible and near-infrared (VNIR) hyperspectral imaging system to collect centimeter-scale data across hypersaline environments. At 30 sites spanning salinity and microhabitat gradients in Great Salt Lake’s North Arm, including Rozel Point and an industrial pond complex, aerial imagery will be paired with shotgun metagenomic sequencing and high-performance liquid chromatography (HPLC) pigment measurements. These data will be used to develop and validate models that estimate bacterioruberin, bacteriorhodopsin, and accessory carotenoid abundance from spectral imagery. Finally, we will create an open-source machine-learning platform that integrates spectral, genomic, pigment, and environmental data to map microbial biogeography and evaluate whether taxonomic composition and functional genes can be predicted from hyperspectral measurements.