Surveillance of pathogenic bacteria on a food matrix using machine-learning-enabled paper chromogenic arrays.

Global food systems can benefit significantly from continuous monitoring of microbial food safety, a task for which tedious operations, destructive sampling, and the inability to monitor multiple pathogens remain challenging. This study reports significant improvements to a paper chromogenic array sensor - machine learning (PCA-ML) methodology sensing concentrations of volatile organic compounds (VOCs) emitted on a species-specific basis by pathogens by streamlining dye selection, sensor fabrication, database construction, and machine learning and validation. This approach enables noncontact, time-dependent, simultaneous monitoring of multiple pathogens (Listeria monocytogenes, Salmonella, and E. coli O157:H7) at levels as low as 1 log CFU/g with over 90% accuracy. The report provides theoretical and practical frameworks demonstrating that chromogenic response, including limits of detection, depends on time integrals of VOC concentrations. The paper also discusses the potential for implementing PCA-ML in the food supply chain for different food matrices and pathogens, with species- and strain-specific identification.

Temporally quasiperiodic data, propagating in the laboratory frame, can be rendered periodic by Galilean transformation.

For a broad class of distributions of temperature, concentration, or another quantity propagating rectilinearly, we show that temporally quasiperiodic behavior in the laboratory frame can be rendered periodic by Galilean transformation. The approach is illustrated analytically and numerically using as an example a closed-form model distribution generated from a one-dimensional partial differential equation, and a detailed process is developed to determine frame speed from more general quasiperiodic, one-dimensional, temporally- and spatially-discretized data. The approach is extended to two- and three-dimensional rectilinear propagation, and its application to nonrectilinear propagation, along with implications for interpreting noise-corrupted data, are also discussed.

Nondestructive multiplex detection of foodborne pathogens with background microflora and symbiosis using a paper chromogenic array and advanced neural network.

We have developed an inexpensive, standardized paper chromogenic array (PCA) integrated with a machine learning approach to accurately identify single pathogens (Listeria monocytogenes, Salmonella Enteritidis, or Escherichia coli O157:H7) or multiple pathogens (either in multiple monocultures, or in a single cocktail culture), in the presence of background microflora on food. Cantaloupe, a commodity with significant volatile organic compound (VOC) emission and large diverse populations of background microflora, was used as the model food. The PCA was fabricated from a paper microarray via photolithography and paper microfluidics, into which 22 chromogenic dye spots were infused and to which three red/green/blue color-standard dots were taped. When exposed to VOCs emitted by pathogens of interest, dye spots exhibited distinguishable color changes and pattern shifts, which were automatically segmented and digitized into a ΔR/ΔG/ΔB database. We developed an advanced deep feedforward neural network with a learning rate scheduler, L2 regularization, and shortcut connections. After training on the ΔR/ΔG/ΔB database, the network demonstrated excellent performance in identifying pathogens in single monocultures, multiple monocultures, and in cocktail culture, and in distinguishing them from the background signal on cantaloupe, providing accuracy of up to 93% and 91% under ambient and refrigerated conditions, respectively. With its combination of speed, reliability, portability, and low cost, this nondestructive approach holds great potential to significantly advance culture-free pathogen detection and identification on food, and is readily extendable to other food commodities with complex microflora.

Machine learning-enabled non-destructive paper chromogenic array detection of multiplexed viable pathogens on food.

Fast and simultaneous identification of multiple viable pathogens on food is critical to public health. Here we report a pathogen identification system using a paper chromogenic array (PCA) enabled by machine learning. The PCA consists of a paper substrate impregnated with 23 chromogenic dyes and dye combinations, which undergo colour changes on exposure to volatile organic compounds emitted by pathogens of interest. These colour changes are digitized and used to train a multi-layer neural network (NN), endowing it with high-accuracy (91-95%) strain-specific pathogen identification and quantification capabilities. The trained PCA-NN system can distinguish between viable Escherichia coli, E. coli O157:H7 and other viable pathogens, and can simultaneously identify both E. coli O157:H7 and Listeria monocytogenes on fresh-cut romaine lettuce, which represents a realistic and complex environment. This approach has the potential to advance non-destructive pathogen detection and identification on food, without enrichment, culturing, incubation or other sample preparation steps.

Fabrication of biomimetically patterned surfaces and their application to probing plant-bacteria interactions.

We have developed a two-step replica molding method for rapid fabrication of biomimetically patterned plant surfaces (BPS) using polydimethylsiloxane (PDMS-BPS) and agarose (AGAR-BPS). Beyond providing multiple identical specimens that faithfully reproduce leaf surface microstructure, this approach also offers unique chemical, physical, and biological features. PDMS-BPS provide good structural durability for SEM examination, have surface wettability comparable to plant surfaces for coating development, and allow for real-time monitoring of biosynthesis through incorporation into microfluidic devices. AGAR-BPS are compatible with bacterial growth, recovery, and quantification, and enable investigation of the effects of surface topography on spatially varying survival and inactivation of Escherichia coli cells during biocide treatment. Further development and application of these biomimetically patterned surfaces to study (and possibly modify) other aspects of plant-bacteria interactions can provide insight into controlling pathogen contamination in a wide range of applications.

Whole-head washing, prior to cutting, provides sanitization advantages for fresh-cut Iceberg lettuce (Latuca sativa L.).

The efficacy of two leafy produce wash methods, the traditional cutting-before-washing process and a new washing-before-cutting method, on reduction of Escherichia coli O157:H7 inoculated on Iceberg lettuce was compared. The washing tests were conducted in a pilot-scale washer using combinations of water, chlorine, peroxyacetic acid, and ultrasound. The washing-before-cutting process recorded an E. coli O157:H7 count reduction 0.79-0.80 log10 CFU/g higher than that achieved with the cutting-before-washing process in treatments involving only a sanitizer. When ultrasound was applied to the washing-before-cutting process, a further improvement of 0.37-0.68 log10 CFU/g in microbial count reduction was obtained, reaching total reductions of 2.43 and 2.24 log10 CFU/g for chlorine and peroxyacetic acid washes, respectively.

The effect of salt on protein chemical potential determined by ternary diffusion in aqueous solutions.

We use accurate thermodynamic derivatives extracted from high-precision measurements of the four volume-fixed diffusion coefficients in ternary solutions of lysozyme chloride in aqueous NaCl, NH4Cl, and KCl at pH 4.5 and 25 degrees C to (a) assess the relative contributions of the common-ion and nonideality effects to the protein chemical potential as a function of salt concentration, (b) compare the behavior of the protein chemical potential for the three salts, which we found to be consistent with the Hofmeister series, and (c) discuss our thermodynamic data in relation to the dependence of the protein solubility on salt concentration. The four diffusion coefficients are reported at 0.6 mM lysozyme chloride and 0.25, 0.5, 0.9, 1.2, and 1.5 M KCl and extend into the protein-supersaturated region. The chemical potential cross-derivatives are extracted from diffusion data using the Onsager reciprocal relation and the equality of molal cross-derivatives of solute chemical potentials. They are compared to those calculated previously from diffusion data for lysozyme in aqueous NaCl and NH4Cl. We estimate the effective charge on the diffusing lysozyme cation at the experimental concentrations. Our diffusion measurements on the three salts allowed us to analyze and interpret the four diffusion coefficients for charged proteins in the presence of 1:1 electrolytes. Our results may provide guidance to the understanding of protein crystallization.