[Nature Sensors] Deep sea microorganisms remain largely inaccessible because existing sampling and cultivation methods fail to preserve native high-pressure microenvironments from seafloor to laboratory.
Here we show a closed-loop cyber physical platform that maintains in situ deep-sea conditions through real-time sensing, digital twin control and robotic high-pressure manipulation to enable targeted isolation of extremophiles.
The system integrates intelligent site selection, active pressure retention and multimodal vision–tactile feedback to automate colony recognition with 94.6% accuracy and perform compliant streaking and picking, enabling the recovery of deep-sea-adapted strains from cold seep environments.
These results establish a sensor-driven framework for standardized exploration of the deep biosphere and provide a generalizable approach for isolating microorganisms from environments inaccessible to conventional methods.
a, The sensing–actuation gap. Conventional ‘non-fidelity’ sampling methods sever the link between the organism and its native environment. The governing metabolic rate equation (inset) illustrates how stochastic fluctuations in substrate (S), microbial abundance (M), temperature (T) and pressure (P) accelerate the decay coefficient (k), leading to the loss of physiological activity (‘decompression transients’). Eα, activation energy; R, ideal gas constant; k0, baseline metabolic rate; S1 to Sn, multiple substrate concentrations; M1 to Mm, multiple microbial abundance; Popt, optimum pressure; 𝜎, pressure tolerance; Yield, growth yield per metabolic rate 𝑘mortality, mortality rate per biomass. b, Closed-loop control hierarchy. The DISCUSS platform bridges this gap via a two-layer cyber-physical architecture. Top (physical layer): an integrated workflow linking high-pressure sampling, enrichment and robotic isolation. The ‘Digital Twin’ module captures in situ parameters to guide the ‘Adaptive Fidelity Transfer’ protocol, ensuring seamless isobaric handover between devices. Bottom (cyber layer): the central control unit (CCU) ingests real-time streams from the multimodal sensor array (visual, haptic, electrochemical and physical). These inputs drive the model predictive control (MPC) and decision-making algorithms, which actively modulate the actuation interface (pumps, thermal regulators and robotic arm) to reconstruct the deep-sea microenvironment with high precision. — [Nature Sensors]
Closed-loop in situ isolation of deep-sea extremophiles through sensor-driven microenvironment preservation, Nature Sensors (open access)
Astrobiology
