Classification of foodborne pathogens using near infrared (NIR) laser scatter imaging system with multivariate calibration

Identifying antibiotic-resistant strains via cell sorting and elastic-light-scatter phenotyping

The proliferation and dissemination of antimicrobial-resistant bacteria is an increasingly global challenge and is attributed mainly to the excessive or improper use of antibiotics. Currently, the gold-standard phenotypic methodology for detecting resistant strains is agar plating, which is a time-consuming process that involves multiple subculturing steps. Genotypic analysis techniques are fast, but they require pure starting samples and cannot differentiate between viable and non-viable organisms. Thus, there is a need to develop a better method to identify and prevent the spread of antimicrobial resistance. This work presents a novel method for detecting and identifying antibiotic-resistant strains by combining a cell sorter for bacterial detection and an elastic-light-scattering method for bacterial classification. The cell sorter was equipped with safety mechanisms for handling pathogenic organisms and enabled precise placement of individual bacteria onto an agar plate. The patterning was performed on an antibiotic-gradient plate, where the growth of colonies in sections with high antibiotic concentrations confirmed the presence of a resistant strain. The antibiotic-gradient plate was also tested with an elastic-light-scattering device where each colony's unique colony scatter pattern was recorded and classified using machine learning for rapid identification of bacteria. Sorting and patterning bacteria on an antibiotic-gradient plate using a cell sorter reduced the number of subculturing steps and allowed direct qualitative binary detection of resistant strains. Elastic-light-scattering technology is a rapid, label-free, and non-destructive method that permits instantaneous classification of pathogenic strains based on the unique bacterial colony scatter pattern. KEY POINTS: • Individual bacteria cells are placed on gradient agar plates by a cell sorter • Laser-light scatter patterns are used to recognize antibiotic-resistant organisms • Scatter patterns formed by colonies correspond to AMR-associated phenotypes.

Rapid and sensitive detection of Salmonella in agro-Food and environmental samples: A review of advances in rapid tests and biosensors

Salmonella is as an intracellular bacterium, causing many human fatalities when the host-specific serotypes reach the host gastrointestinal tract. Nontyphoidal Salmonella are responsible for numerous foodborne outbreaks and product recalls worldwide whereas typhoidal Salmonella are responsible for Typhoid fever cases in developing countries. Yet, Salmonella-related foodborne disease outbreaks through its food and water contaminations have urged the advancement of rapid and sensitive Salmonella-detecting methods for public health protection. While conventional detection methods are time-consuming and ineffective for monitoring foodstuffs with short shelf lives, advances in microbiology, molecular biology and biosensor methods have hastened the detection. Here, the review discusses Salmonella pathogenic mechanisms and its detection technology advancements (fundamental concepts, features, implementations, efficiency, benefits, limitations and prospects). The time-efficiency of each rapid test method is discussed in relation to their limit of detections (LODs) and time required from sample enrichment to final data analysis. Importantly, the matrix effects (LODs and sample enrichments) were compared within the methods to potentially speculate Salmonella detection from environmental, clinical or food matrices using certain techniques. Although biotechnological advancements have led to various time-efficient Salmonella-detecting techniques, one should consider the usage of sophisticated equipment to run the analysis by moderately to highly trained personnel. Ultimately, a fast, accurate Salmonella screening that is readily executed by untrained personnels from various matrices, is desired for public health procurement.

How to Resolve the Maximum Valuable Information in Complex NIR Signal: A Practicable Method Based on Wavelet Transform

A key problem in the field of near infrared (NIR) spectrum study is to obtain the valuable information from the complex NIR signal. A maximum information extraction method based on Wavelet Transform (WT) is proposed in this paper for helping the relative researchers to resolve the signal. The results show that the method can serve as an effective tool for obtaining the maximum valuable information in NIR study.

Development of a Smartphone-Integrated Reflective Scatterometer for Bacterial Identification

We present a smartphone-based bacterial colony phenotyping instrument using a reflective elastic light scattering (ELS) pattern and the resolving power of the new instrument. The reflectance-type device can acquire ELS patterns of colonies on highly opaque media as well as optically dense colonies. The novel instrument was built using a smartphone interface and a 532 nm diode laser, and these essential optical components made it a cost-effective and portable device. When a coherent and collimated light source illuminated a bacterial colony, a reflective ELS pattern was created on the screen and captured by the smartphone camera. The collected patterns whose shapes were determined by the colony morphology were then processed and analyzed to extract distinctive features for bacterial identification. For validation purposes, the reflective ELS patterns of five bacteria grown on opaque growth media were measured with the proposed instrument and utilized for the classification. Cross-validation was performed to evaluate the classification, and the result showed an accuracy above 94% for differentiating colonies of E. coli, K. pneumoniae, L. innocua, S. enteritidis, and S. aureus.

Foodborne Disease Risk Prediction Using Multigraph Structural Long Short-term Memory Networks: Algorithm Design and Validation Study

BACKGROUND

Foodborne disease is a common threat to human health worldwide, leading to millions of deaths every year. Thus, the accurate prediction foodborne disease risk is very urgent and of great importance for public health management.

OBJECTIVE

We aimed to design a spatial-temporal risk prediction model suitable for predicting foodborne disease risks in various regions, to provide guidance for the prevention and control of foodborne diseases.

METHODS

We designed a novel end-to-end framework to predict foodborne disease risk by using a multigraph structural long short-term memory neural network, which can utilize an encoder-decoder to achieve multistep prediction. In particular, to capture multiple spatial correlations, we divided regions by administrative area and constructed adjacent graphs with metrics that included region proximity, historical data similarity, regional function similarity, and exposure food similarity. We also integrated an attention mechanism in both spatial and temporal dimensions, as well as external factors, to refine prediction accuracy. We validated our model with a long-term real-world foodborne disease data set, comprising data from 2015 to 2019 from multiple provinces in China.

RESULTS

Our model can achieve F1 scores of 0.822, 0.679, 0.709, and 0.720 for single-month forecasts for the provinces of Beijing, Zhejiang, Shanxi and Hebei, respectively, and the highest F1 score was 20% higher than the best results of the other models. The experimental results clearly demonstrated that our approach can outperform other state-of-the-art models, with a margin.

CONCLUSIONS

The spatial-temporal risk prediction model can take into account the spatial-temporal characteristics of foodborne disease data and accurately determine future disease spatial-temporal risks, thereby providing support for the prevention and risk assessment of foodborne disease.

Optical multi-channel interrogation instrument for bacterial colony characterization

A single instrument that includes multiple optical channels was developed to simultaneously measure various optical and associated biophysical characteristics of a bacterial colony. The multi-channel device can provide five distinct optical features without the need to transfer the sample to multiple locations or instruments. The available measurement channels are bright-field light microscopy, 3-D colony-morphology map, 2-D spatial optical-density distribution, spectral forward-scattering pattern, and spectral optical density. The series of multiple morphological interrogations is beneficial in understanding the bio-optical features of a bacterial colony and the correlations among them, resulting in an enhanced power of phenotypic bacterial discrimination. To enable a one-shot interrogation, a confocal laser scanning module was built as an add-on to an upright microscope. Three different-wavelength diode lasers were used for the spectral analysis, and high-speed pin photodiodes and CMOS sensors were utilized as detectors to measure the spectral OD and light-scatter pattern. The proposed instrument and algorithms were evaluated with four bacterial genera, Escherichia coli, Listeria innocua, Salmonella Typhimurium, and Staphylococcus aureus; their resulting data provided a more complete picture of the optical characterization of bacterial colonies.

Machine Learning Prediction of Foodborne Disease Pathogens: Algorithm Development and Validation Study

Background

Foodborne diseases, as a type of disease with a high global incidence, place a heavy burden on public health and social economy. Foodborne pathogens, as the main factor of foodborne diseases, play an important role in the treatment and prevention of foodborne diseases; however, foodborne diseases caused by different pathogens lack specificity in clinical features, and there is a low proportion of clinically actual pathogen detection in real life.

Objective

We aimed to analyze foodborne disease case data, select appropriate features based on analysis results, and use machine learning methods to classify foodborne disease pathogens to predict foodborne disease pathogens that have not been tested.

Methods

We extracted features such as space, time, and exposed food from foodborne disease case data and analyzed the relationship between these features and the foodborne disease pathogens using a variety of machine learning methods to classify foodborne disease pathogens. We compared the results of 4 models to obtain the pathogen prediction model with the highest accuracy.

Results

The gradient boost decision tree model obtained the highest accuracy, with accuracy approaching 69% in identifying 4 pathogens including Salmonella, Norovirus, Escherichia coli, and Vibrio parahaemolyticus. By evaluating the importance of features such as time of illness, geographical longitude and latitude, and diarrhea frequency, we found that they play important roles in classifying the foodborne disease pathogens.

Conclusions

Data analysis can reflect the distribution of some features of foodborne diseases and the relationship among the features. The classification of pathogens based on the analysis results and machine learning methods can provide beneficial support for clinical auxiliary diagnosis and treatment of foodborne diseases.

Generalized spectral light scatter models of diverse bacterial colony morphologies

An optical forward-scatter model was generalized to encompass the diverse nature of bacterial colony morphologies and the spectral information. According to the model, the colony shape and the wavelength of incident light significantly affect the characteristics of a forward elastic-light-scattering pattern. To study the relationship between the colony morphology and the scattering pattern, three-dimensional colony models were generated in various morphologies. The propagation of light passing through the colony model was then simulated. In validation of the theoretical modeling, the scattering patterns of three bacterial genera, Staphylococcus, Exiguobacterium and Bacillus, which grow into colonies having convex, crateriform and flat elevations, respectively, were qualitatively compared to the simulated scattering patterns. The strong correlations observed between simulated and experimental patterns validated the scatter model. In addition, spectral effect on the scattering pattern was studied using the scatter model, and experimentally investigated using Staphylococcus, whose colony has circular form and convex elevation. Both simulation and experiment showed that changes in wavelength affected the overall pattern size and the number of rings.

Disposable all-printed electronic biosensor for instantaneous detection and classification of pathogens

A novel disposable all-printed electronic biosensor is proposed for a fast detection and classification of bacteria. This biosensor is applied to classify three types of popular pathogens: Salmonella typhimurium, and the Escherichia coli strains JM109 and DH5-α. The proposed sensor consists of inter-digital silver electrodes fabricated through an inkjet material printer and silver nanowires uniformly decorated on the electrodes through the electrohydrodynamic technique on a polyamide based polyethylene terephthalate substrate. The best sensitivity of the proposed sensor is achieved at 200 µm teeth spaces of the inter-digital electrodes along the density of the silver nanowires at 30 × 10³/mm². The biosensor operates on ±2.5 V and gives the impedance value against each bacteria type in 8 min after sample injection. The sample data are measured through an impedance analyzer and analyzed through pattern recognition methods such as linear discriminate analysis, maximum likelihood, and back propagation artificial neural network to classify each type of bacteria. A perfect classification and cross-validation is achieved by using the unique fingerprints extracted from the proposed biosensor through all the applied classifiers. The overall experimental results demonstrate that the proposed disposable all-printed biosensor is applicable for the rapid detection and classification of pathogens.

Development of a multispectral light-scatter sensor for bacterial colonies

We report a multispectral elastic-light-scatter instrument that can simultaneously detect three-wavelength scatter patterns and associated optical densities from individual bacterial colonies, overcoming the limits of the single-wavelength predecessor. Absorption measurements on liquid bacterial samples revealed that the spectroscopic information can indeed contribute to sample differentiability. New optical components, including a pellicle beam splitter and an optical cage system, were utilized for robust acquisition of multispectral images. Four different genera and seven shiga toxin producing E. coli serovars were analyzed; the acquired images showed differences in scattering characteristics among the tested organisms. In addition, colony-based spectral optical-density information was also collected. The optical model, which was developed using diffraction theory, correctly predicted wavelength-related differences in scatter patterns, and was matched with the experimental results. Scatter-pattern classification was performed using pseudo-Zernike (GPZ) polynomials/moments by combining the features collected at all three wavelengths and selecting the best features via a random-forest method. The data demonstrate that the selected features provide better classification rates than the same number of features from any single wavelength. Three wavelength-merged scatter pattern from E. coli.

Early detection of germinated wheat grains using terahertz image and chemometrics

In this paper, we propose a feasible tool that uses a terahertz (THz) imaging system for identifying wheat grains at different stages of germination. The THz spectra of the main changed components of wheat grains, maltose and starch, which were obtained by THz time spectroscopy, were distinctly different. Used for original data compression and feature extraction, principal component analysis (PCA) revealed the changes that occurred in the inner chemical structure during germination. Two thresholds, one indicating the start of the release of α-amylase and the second when it reaches the steady state, were obtained through the first five score images. Thus, the first five PCs were input for the partial least-squares regression (PLSR), least-squares support vector machine (LS-SVM), and back-propagation neural network (BPNN) models, which were used to classify seven different germination times between 0 and 48 h, with a prediction accuracy of 92.85%, 93.57%, and 90.71%, respectively. The experimental results indicated that the combination of THz imaging technology and chemometrics could be a new effective way to discriminate wheat grains at the early germination stage of approximately 6 h.

Fast quantifying collision strength index of ethylene-vinyl acetate copolymer coverings on the fields based on near infrared hyperspectral imaging techniques

A novel strategy based on the near infrared hyperspectral imaging techniques and chemometrics were explored for fast quantifying the collision strength index of ethylene-vinyl acetate copolymer (EVAC) coverings on the fields. The reflectance spectral data of EVAC coverings was obtained by using the near infrared hyperspectral meter. The collision analysis equipment was employed to measure the collision intensity of EVAC materials. The preprocessing algorithms were firstly performed before the calibration. The algorithms of random frog and successive projection (SP) were applied to extracting the fingerprint wavebands. A correlation model between the significant spectral curves which reflected the cross-linking attributions of the inner organic molecules and the degree of collision strength was set up by taking advantage of the support vector machine regression (SVMR) approach. The SP-SVMR model attained the residual predictive deviation of 3.074, the square of percentage of correlation coefficient of 93.48% and 93.05% and the root mean square error of 1.963 and 2.091 for the calibration and validation sets, respectively, which exhibited the best forecast performance. The results indicated that the approaches of integrating the near infrared hyperspectral imaging techniques with the chemometrics could be utilized to rapidly determine the degree of collision strength of EVAC.

Fabricating Upconversion Fluorescent Probes for Rapidly Sensing Foodborne Pathogens

Rare earth-doped upconversion nanoparticles (UCNPs) have promising potential in the field of food safety because of their unique frequency upconverting capability and high detection sensitivity. Here, we report a rapid and sensitive UCNP-based bacterium-sensing strategy using Escherichia coli. Highly fluorescent and water-soluble UCNPs were fabricated and conjugated with antibodies against E. coli for use as fluorescent probes. The E. coli were successively captured by the fluorescent probes. After the captured cell samples were pelleted, the differences in the fluorescence intensities between sample supernatants and the control were observed to increase linearly with E. coli concentration from 42 to 42 × 10(6) colony-forming units (cfu)/mL (R(2) = 0.9802), resulting in a relatively low limit of detection of 10 cfu/mL. Furthermore, the ability of the bioassay to detect E. coli was also confirmed in adulterated meat and milk samples.