AI-enabled biosensing for rapid pathogen detection: From liquid food to agricultural water.

Rapid pathogen detection in food and agricultural water is essential for ensuring food safety and public health. However, complex and noisy environmental background matrices delay the identification of pathogens and require highly trained personnel. Here, we present an AI-biosensing framework for accelerated and automated pathogen detection in various water samples, from liquid food to agricultural water. A deep learning model was used to identify and quantify target bacteria based on their microscopic patterns generated by specific interactions with bacteriophages. The model was trained on augmented datasets to maximize data efficiency, using input images of selected bacterial species, and then fine-tuned on a mixed culture. Model inference was performed on real-world water samples containing environmental noises unseen during model training. Overall, our AI model trained solely on lab-cultured bacteria achieved rapid (< 5.5 h) prediction with 80-100% accuracy on the real-world water samples, demonstrating its ability to generalize to unseen data. Our study highlights the potential applications in microbial water quality monitoring during food and agricultural processes.

Design and Development of Robust, Daylight-Activated, and Rechargeable Biocidal Polymeric Films as Promising Active Food Packaging Materials.

The emerging infectious diseases have created one of the major practical needs to develop active packaging materials with durable antibacterial and antiviral properties for the food industry. To meet this demand, the development of new technologies applicable to food contact surfaces is highly desired but challenging. The recent discovery of the photoactive properties of vitamin K (VK) derivatives has raised great expectations as promising candidates in functional film development due to the generation of biocidal reactive oxygen species (ROS) by these compounds. Inspired by the excellent photoactivity of one of the light-stable VK derivatives, menadione (VK3), under visible daylight irradiation, we demonstrate a protocol for the fabrication of daylight-mediated biocidal packaging materials by incorporating VK3 into a poly (ethylene-co-vinyl acetate) (EVA) matrix. The VK3 (i.e., 1-5% w/w) incorporated EVA films successfully demonstrated the production of ROS and antibacterial and antiviral performance against Escherichia coli, Listeria innocua, and T7 bacteriophage, respectively, under daylight exposure conditions. The results revealed that the addition of a proper percentage of VK3 significantly enhanced the ROS productivity of the films and created a novel daylight-induced microbial killing performance on the films. The biocidal functions of the films are long-lasting and rechargeable when exposed to light repeatedly, making them a viable contender for replacing currently available conventional packaging films.

Accelerating the Detection of Bacteria in Food Using Artificial Intelligence and Optical Imaging.

In assessing food microbial safety, the presence of Escherichia coli is a critical indicator of fecal contamination. However, conventional detection methods require the isolation of bacterial macrocolonies for biochemical or genetic characterization, which takes a few days and is labor-intensive. In this study, we show that the real-time object detection and classification algorithm You Only Look Once version 4 (YOLOv4) can accurately identify the presence of E. coli at the microcolony stage after a 3-h cultivation. Integrating with phase-contrast microscopic imaging, YOLOv4 discriminated E. coli from seven other common foodborne bacterial species with an average precision of 94%. This approach also enabled the rapid quantification of E. coli concentrations over 3 orders of magnitude with an R2 of 0.995. For romaine lettuce spiked with E. coli (10 to 103 CFU/g), the trained YOLOv4 detector had a false-negative rate of less than 10%. This approach accelerates analysis and avoids manual result determination, which has the potential to be applied as a rapid and user-friendly bacterial sensing approach in food industries. IMPORTANCE A simple, cost-effective, and rapid method is desired to identify potential pathogen contamination in food products and thus prevent foodborne illnesses and outbreaks. This study combined artificial intelligence (AI) and optical imaging to detect bacteria at the microcolony stage within 3 h of inoculation. This approach eliminates the need for time-consuming culture-based colony isolation and resource-intensive molecular approaches for bacterial identification. The approach developed in this study is broadly applicable for the identification of diverse bacterial species. In addition, this approach can be implemented in resource-limited areas, as it does not require expensive instruments and significantly trained human resources. This AI-assisted detection not only achieves high accuracy in bacterial classification but also provides the potential for automated bacterial detection, reducing labor workloads in food industries, environmental monitoring, and clinical settings.

Spectroscopy Approaches for Food Safety Applications: Improving Data Efficiency Using Active Learning and Semi-supervised Learning.

The past decade witnessed rapid development in the measurement and monitoring technologies for food science. Among these technologies, spectroscopy has been widely used for the analysis of food quality, safety, and nutritional properties. Due to the complexity of food systems and the lack of comprehensive predictive models, rapid and simple measurements to predict complex properties in food systems are largely missing. Machine Learning (ML) has shown great potential to improve the classification and prediction of these properties. However, the barriers to collecting large datasets for ML applications still persists. In this paper, we explore different approaches of data annotation and model training to improve data efficiency for ML applications. Specifically, we leverage Active Learning (AL) and Semi-Supervised Learning (SSL) and investigate four approaches: baseline passive learning, AL, SSL, and a hybrid of AL and SSL. To evaluate these approaches, we collect two spectroscopy datasets: predicting plasma dosage and detecting foodborne pathogen. Our experimental results show that, compared to the de facto passive learning approach, advanced approaches (AL, SSL, and the hybrid) can greatly reduce the number of labeled samples, with some cases decreasing the number of labeled samples by more than half.

Quantitative Imaging of Bacteriophage Amplification for Rapid Detection of Bacteria in Model Foods.

Rapid detection of bacteria in water and food samples is a critical need. The current molecular methods like real-time PCR can provide rapid detection after initial enrichment. However, these methods require significant preparation steps, specialized facilities to reduce contamination, and relatively expensive reagents. This study evaluates a novel approach for detecting bacteria based on imaging of bacteriophage amplification upon infection of the target host bacteria to mitigate some of these constraints and improve the specificity of discriminating live vs. dead bacteria. Thus, this research leverages the natural ability of lytic bacteriophages to rapidly amplify their genetic material and generate progeny phages upon infecting the host bacterium. This study uses a nucleic acid staining dye, a conventional fluorescence microscope, and quantitative image analysis for imaging the amplification of bacteriophages. The sensitivity and assay time for imaging-based quantification of phage amplification for detecting Escherichia coli were compared with RT-PCR and the standard plaque-forming assay for detection phage amplification in model systems, including coconut water and spinach wash water. The results demonstrate that the imaging approach matches both the sensitivity and speed for detecting E. coli using the RT-PCR method without requiring isolation of nucleic acids, expensive reagents, and specialized facilities. The quantitative imaging results demonstrate the detection of 10 CFU/ml of E. coli in coconut water and simulated spinach wash water with a chemical oxygen demand (COD) of 3,000 ppm within 8 h, including initial enrichment of the bacteria. In summary, the results of this study illustrate a novel phage amplification-based approach for detecting target bacteria in complex food and water samples using simple sample preparation methods and low-cost reagents.

Application of Engineered Bacteriophage T7 in the Detection of Bacteria in Food Matrices.

Detection of pathogens in a food matrix is challenging due to various constraints including complexity and the cost of sample preparation for microbial analysis from food samples, time period for the detection of pathogens, and high cost and specialized resources required for advanced molecular assays. To address some of these key challenges, this study illustrates a simple and rapid colorimetric detection of target bacteria in distinct food matrices, including fresh produce, without prior isolation of bacteria from a food matrix. This approach combines bacteriophage-induced expression of an exogenous enzyme, alkaline phosphatase, the specific colorimetric substrate that generates insoluble color products, and a simple filtration method to localize the generation of colored signal. Using this approach, this study demonstrates the specific detection of inoculated Escherichia coli in coconut water and baby spinach leaves. Without isolating bacteria from the selected food matrices and using a food sample size that is representative of industrial samples, the inoculated samples were added to the enrichment broth for a short period (5 h) and incubated with an engineered bacteriophage T7 with a phoA gene. The incubation period with the engineered bacteriophage was 30 min for liquid samples and 2 h for fresh produce samples. The samples were then filtered through a 0.2-micron polycarbonate membrane and incubated with a colorimetric substrate, i.e., nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP). This substrate forms a dark purple precipitate upon interactions with the released enzyme on a filter membrane. This approach successfully detected 10 CFU/ml of E. coli in coconut water and 102 CFU/g of E. coli on baby spinach leaves with 5 h of enrichment. Success of this approach illustrates potential for detecting target bacteria in food systems using a simple visual assay and/or quantitative colorimetric measurements.

A Novel N-Halamine Biocidal Nanofibrous Membrane for Chlorine Rechargeable Rapid Water Disinfection Applications.

Disinfecting pathogenic contaminated water rapidly and effectively on sites is one of the critical challenges at point-of-use (POU) situations. Currently available technologies are still suffering from irreversible depletion of disinfectants, generation of toxic by-products, and potential biofouling problems. Herein, we developed a chlorine rechargeable biocidal nanofibrous membrane, poly(acrylonitrile-co-5-methyl-5-(4'-vinylphenyl)imidazolidine-2,4-dione) (P(AN-VAPH)), via a combination of a free radical copolymerization reaction and electrospun technology. The copolymer exhibits good electrospinnability and desirable mechanical properties. Also, the 5-methyl-5-(4'-vinylphenyl)imidazolidine-2,4-dione (VAPH) moieties containing unique hydantoin structures are able to be chlorinated and converted to halamine structures, enabling the P(AN-VAPH) nanofibrous membrane with rapid and durable biocidal activity. The chlorinated P(AN-VAPH) nanofibrous membranes showed intriguing features of unique 3D morphological structures with large specific surface area, good mechanical performance, rechargeable chlorination capacity (>5000 ppm), long-term durability, and desirable biocidal activity against both bacteria and viruses (>99.9999% within 2 min of contact). With these attributes, the chlorinated P(AN-VAPH) membranes demonstrated promising disinfecting efficiency against concentrated bacteria-contaminated water during direct filtration applications with superior killing capacity and high flowing flux (5000 L m-2 h-1).

N-Halamine Polypropylene Nonwoven Fabrics with Rechargeable Antibacterial and Antiviral Functions for Medical Applications.

Embedding medical and hygiene products with regenerable antimicrobial functions would have significant implications for limiting pathogen contaminations and reducing healthcare-associated infections. Herein, we demonstrate a scalable and industrially feasible methodology to fabricate chlorine rechargeable melt-blown polypropylene (PP) nonwoven fabrics, which have been widely used in hygienic and personal protective products, via a combination of a melt reactive extrusion process and melt-blown technique. Methacrylamide (MAM) was employed as a precursor of halamine monomers and covalently grafted onto the PP backbone to form polypropylene-grafted methacrylamide (PP-g-MAM), which could be chlorinated, yielding biocidal acyclic halamines. Subsequently, the resultant PP-g-MAM was manufactured into nonwoven fabrics with varying fiber diameters by adjusting the hot air flowing speed during the melt-blowing process. The chlorinated nonwoven fabrics (PP-g-MAM-Cl) exhibited integrated properties such as a robust mechanical property, good thermal stability, high chlorination capability (>850 ppm), and desirable chlorine rechargeability. More importantly, such chlorinated nonwoven fabrics showed a promising antibacterial and antiviral efficiency, achieving 6 log CFU reduction of bacteria (both Escherichia coliO157: H7 and Listeria innocua) and 7 log PFU reductions of a virus (T7 bacteriophages) within 15 and 5 min of contact, respectively, revealing great potential to serve as a reusable antimicrobial material for medical protection applications.

Daylight-Induced Antibacterial and Antiviral Nanofibrous Membranes Containing Vitamin K Derivatives for Personal Protective Equipment.

During the development of antibacterial and antiviral materials for personal protective equipment (PPE), daylight active functional polymeric materials containing vitamin K compounds (VKs) and impacts of polymer structures to the functions were investigated. As examples, hydrophobic polyacrylonitrile (PAN) and hydrophilic poly(vinyl alcohol-co-ethylene) (PVA-co-PE) polymers were directly blended with three VK compounds and electrospun into VK-containing nanofibrous membranes (VNFMs). The prepared VNFMs exhibited robust photoactivity in generating reactive oxygen species (ROS) under both daylight (D65, 300-800 nm) and ultraviolet A (UVA, 365 nm) irradiation, resulting in high antimicrobial and antiviral efficiency (>99.9%) within a short exposure time (<90 min). Interestingly, the PVA-co-PE/VK3 VNFM showed higher ROS production rates and better biocidal functions than those of the PAN/VK3 VNFM under the same photoirradiation conditions, indicating that PVA-co-PE is a better matrix polymer material for these functions. Moreover, the prepared PVA-co-PE/VK3 VNFM maintains its powerful microbicidal function even after five times of repeated exposures to bacteria and viruses, showing the stability and reusability of the antimicrobial materials. The fabrication of photoinduced antimicrobial VNFMs may provide new insights into the development of non-toxic and reusable photoinduced antimicrobial materials that could be applied in personal protective equipment with improved biological protections.

Daylight-Induced Antibacterial and Antiviral Cotton Cloth for Offensive Personal Protection.

Cotton fabrics with durable and reusable daylight-induced antibacterial/antiviral functions were developed by using a novel fabrication process, which employs strong electrostatic interaction between cationic cotton fibers and anionic photosensitizers. The cationic cotton contains polycationic short chains produced by a self-propagation of 2-diehtylaminoehtyl chloride (DEAE-Cl) on the surface of cotton fibers. Then, the fabric (i.e., polyDEAE@cotton) can be readily functionalized with anionic photosensitizers like rose Bengal and sodium 2-anthraquinone sulfate to produce biocidal reactive oxygen species (ROS) under light exposure and consequently provide the photo-induced biocidal functions. The biocidal properties of the photo-induced fabrics (PIFs) were demonstrated by ROS production measurements, bactericidal performance against bacteria (e.g., E coli and L. innocua), and antiviral results against T7 bacteriophage. The PIFs achieved 99.9999% (6 log) reductions against bacteria and the bacteriophage within 60 min of daylight exposure. Moreover, the PIFs showcase excellent washability and photostability, making them ideal materials for reusable face masks and protective suits with improved biological protections compared with traditional PPE. This work demonstrated that the cationized cotton could serve as a platform for different functionalization applications, and the resulting fiber materials could inspire the development of reusable and sustainable PPE with significant bioprotective properties to fight the COVID-19 pandemic as well as the spread of other contagious diseases.

Rapid detection of Escherichia coli using bacteriophage-induced lysis and image analysis.

Rapid detection of bacterial pathogens is a critical unmet need for both food and environmental samples such as irrigation water. As a part of the Food safety Modernization Act (FSMA), The Produce Safety rule has established several requirements for testing for the presence of generic Escherichia coli in water, but the current method available for testing (EPA M1603) demands specified multiple colony verification and highly trained personnel to perform these tests. The purpose of the study was to assess a phage induced bacterial lysis using quantitative image analysis to achieve rapid detection of E. coli at low concentrations within 8 hours. This study aimed to develop a simple yet highly sensitive and specific approach to detect target bacteria in complex matrices. In the study, E. coli cells were first enriched in tryptic soy broth (TSB), followed by T7 phage induced lysis, concentration, staining and fluorescent imaging. Image analysis was conducted including image pre-processing, image segmentation and quantitatively analysis of cellular morphological features (area, eccentricity and full width at half maximum). Challenge experiments using realistic matrices, including simulated fresh produce wash water, coconut water and spinach wash water, demonstrated the method can be applied for use in situations that occur in food processing facilities. The results indicated E. coli cells that are lysed by T7 phages demonstrated significantly (P < 0.05) higher extracellular DNA release, altered cellular shape (from rod to circular) and diffused fluorescent signal intensity. Using this biosensing strategy, a sensitivity to detect Escherichia coli at 10 CFU/ml within 8 hours was achieved, both in laboratory medium and in complex matrices. The proposed phage based biosensing strategy enables rapid detection of bacteria and is applicable to analysis of food systems. Furthermore, the steps involved in this assay can be automated to enable detection of target bacteria in food facilities without extensive resources.

Rapid detection of Escherichia coli in beverages using genetically engineered bacteriophage T7.

Foodborne illness due to bacterial contamination is a significant issue impacting public health that demands new technology which is practical to implement by food industry. Detection of bacteria in food products and production facilities is a crucial strategy supporting food safety assessments. Bacteriophages were investigated as a tool for bacterial detection due to their ability to infect specific strain of host bacteria in order to improve sensitivity, specificity, and rapidity of bacterial detection. The results of this investigation reveal a novel method for rapid detection. The method employs a genetically engineered bacteriophage, phage T7-ALP, which expresses alkaline phosphatase. Upon infection of Escherichia coli, overexpression of alkaline phosphatase provides an opportunity for rapid sensitive detection of a signal indicative of bacterial presence in model beverage samples as low as 100 bacteria per gram. The method employs a fluorescent precipitated substrate, ELF-97, as a substrate for alkaline phosphatase activity coupled with fluorescence imaging and image analysis allowing single-cell imaging results in high detection sensitivity. The method is easily completed within less than 6 h enabling it to be deployed within most large industrial food processing facilities that have routine 8-h operational shifts.