Progress 03/01/22 to 02/28/23
Outputs
Target Audience:The results of this project address the unmet needs for rapid and specific detection of low numbers of viable pathogens in food and environmental samples within the agriculture and food processing industries. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?This project has supported the research training of a graduate student and two postdoctoral scientists during the past year. These graduate students and postdocs have presented their findings at professional and academic conferences. How have the results been disseminated to communities of interest?The results of this project were disseminated through research presentations at meetings at UC Davis and other institutions What do you plan to do during the next reporting period to accomplish the goals?1. Evaluate the AI-microcolony framework for pathogen detection in mixed microbial cultures containing commensal microbes. 2. Develop more imaging options to increase the adaptability of AI-based framework in different settings of agriculture and food processing industry. In this area, we plan to investigate the phage amplification approach. While phages induce morphological changes in target bacterial cells, they also amplify which can be detected by the increase of signals as a function of time. With a phage-infection time of 1 hour, the progeny phages will be generated and released from the lysed bacterial cells. Since phage particles can be observed using fluorescence DNA staining dye and fluorescence microscopy, phage particles can be enumerated using a simple image analysis. The amplification of the phage particles number indicates the presence of the target bacteria. To apply this method in detection of bacteria in environmental samples with high background noise such as agricultural water, AI approaches may facilitate recognizing phage particles against background noise allowing improved detection sensitivity. 3. Develop a low-cost mobile AI-enabled hardware system and run the model developed in the past year locally without internet transmission to the cloud (i.e., "on the edge").
Impacts
What was accomplished under these goals?
A major study was conducted with the follow goals/outcomes. 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.
Publications
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2022
Citation:
Yi, J., Wisuthiphaet, N., Raja, P., Nitin, N, Earles, M. Combining AI and biosensing for rapid pathogen detection in food
and agricultural water. UC Davis 7th Annual Postdoctoral Research Symposium-Selected Rapid Talk, Davis, CA, Mar 2022.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2022
Citation:
Yi, J., Wisuthiphaet, N., Raja, P., Nitin, N, Earles, M. Combining AI and biosensing for rapid pathogen detection in food and agricultural water. UC Davis 7th Annual Postdoctoral Research Symposium-Selected Rapid Talk, Davis, CA, Mar 2022.
- Type:
Journal Articles
Status:
Published
Year Published:
2023
Citation:
L Ma, J Yi, N Wisuthiphaet, M Earles, N Nitin (2023). Accelerating the detection of bacteria in food using artificial intelligence and optical imaging. Applied and Environmental Microbiology 89 (1), e01828-22
|
Progress 03/01/21 to 02/28/22
Outputs
Target Audience:The results of this project address the unmet needs for rapid and specific detection of low numbers of viable pathogens in food and environmental samples within the agriculture and food processing industries. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?This project has supported the research training of a graduate student and two postdoctoral scientists during the past year. These graduate students and postdocs have presented their findings at professional and academic conferences. How have the results been disseminated to communities of interest?The results of this project were disseminated through research presentations at meetings at UC Davis and other institutions What do you plan to do during the next reporting period to accomplish the goals?1. Evaluate the AI-microcolony framework for pathogen detection in mixed microbial cultures containing commensal microbes. 2. Develop more imaging options to increase the adaptability of AI-based framework in different settings of agriculture and food processing industry. In this area, we plan to investigate the phage amplification approach. While phages induce morphological changes in target bacterial cells, they also amplify which can be detected by the increase of signals as a function of time. With a phage-infection time of 1 hour, the progeny phages will be generated and released from the lysed bacterial cells. Since phage particles can be observed using fluorescence DNA staining dye and fluorescence microscopy, phage particles can be enumerated using a simple image analysis. The amplification of the phage particles number indicates the presence of the target bacteria. To apply this method in detection of bacteria in environmental samples with high background noise such as agricultural water, AI approaches may facilitate recognizing phage particles against background noise allowing improved detection sensitivity. 3. Develop a low-cost mobile AI-enabled hardware system and run the model developed in the past year locally without internet transmission to the cloud (i.e., "on the edge").
Impacts
What was accomplished under these goals?
Detection of pathogenic bacteria in food and environmental samples is vital for ensuing food safety, shelf life, and food security. However, timely detection of pathogenic bacteria is interfered with by complex and noisy environmental background matrices and requires a well-trained workforce. To this end, we developed an AI-biosensing platform for rapid and automated pathogen identification in food and agricultural water. Overall, this AI-biosensing platform can serve as a foundation of automated and accelerated AI-based pathogen detection in agriculture and food processing industries. Obj 1. Develop an AI-based framework for specific detection of viable bacterial pathogens in mixed microbial cultures containing commensal microbes: For this objective, we developed an AI-biosensing platform by training AI model to detect diffusive microscopic patterns of bacteriophage (phage)-induced lysis of target E. coli. We trained a convolutional neural network (CNN) object detection model with fluorescence microscopic images of bacteria labeled with DNA-specific dye and exposed to phages. E. coli and selected non-E. coli bacteria (i.e., Listeria innocua, Bacillus subtilis, Pseudomonas fluorescens) were used to generate training images, where phages induced lysis only in target E. coli cells. The results show that the trained AI model accurately predicted the number and locations of target E. coli cells for three different initial levels of initial loads, i.e., 10, 102, 103 CFU/mL. Moreover, the results show that the trained AI model could accurately detect and quantify E. coli in unseen food and agricultural water samples, including coconut water, spinach wash water, and irrigation water. The prediction accuracy was 80-100% in 403 images, depending on the sources of background noise (i.e., food debris, organic matter, commensal microbes). In addition, this AI-biosensing platform provides a rapid detection with a reduced assay time of 5.5 h, which within a single food processing work shift (4-6 h). The detection limit of this AI-biosensing platform was 10 CFU/mL, which was better or comparable to the sensitivity of other conventional pathogen detection methods such as plate counting assay and real-time quantitative polymerase chain reaction (RT-qPCR), given the samples prepared by 4-h bacterial enrichment. Obj 2. Evaluate point-of-use detection of target pathogens in a simulated limited testing resource environment using automated "edge AI" data acquisition and analysis platform for food and environmental samples: For this objective, we focused on developing a low-cost AI-based framework for pathogen detection that could potentially be applied to the food processing facilities without multiple sample preparation steps. To this end, we developed another AI-based framework where a model was trained on microscopic images of microcolonies (20-50 bacterial cells) without fluorescence labels or phages. Here, 8 distinct bacterial species, including Bacillus coaqulans, Bacillus subtilis, E. coli, Listeria innocua, Listeria monocytogenes, Pseudomonas fluorescens, Salmonella Enteritidis, and Salmonella Typhimurium, were investigated. The microcolonies of these bacterial species were incubated for 3 h prior to imaging, and the acquired images were split into training and test dataset. A CNN image classification model was then trained on training dataset and the prediction was performed on unseen test dataset, resulting in the accuracy of 87-100% across the bacterial species.
Publications
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2021
Citation:
Nitin, N. AI-Enabled Innovations in Food Systems, LLNL Data Science Seminar, 2021
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2021
Citation:
Nitin, N. Innovations in Food Safety with Machine Learning and Novel Antimicrobial Formulations, UIUC, 2022.
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2022
Citation:
Nitin, N. AI-Enabled Innovations in Validation of Sanitation and Detection of Pathogens, AI for food systems in the southeast US, Auburn University, 2022.
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2022
Citation:
Yi, J., Wisuthiphaet, N., Raja, P., Nitin, N, Earles, M. Combining AI and biosensing for rapid pathogen detection in food and agricultural water. UC Davis 7th Annual Postdoctoral Research Symposium-Poster Presentation, Davis, CA, Mar 2022.
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2022
Citation:
Yi, J., Wisuthiphaet, N., Raja, P., Nitin, N, Earles, M. Combining AI and biosensing for rapid pathogen detection in food and agricultural water. UC Davis 7th Annual Postdoctoral Research Symposium-Selected Rapid Talk, Davis, CA, Mar 2022.
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2022
Citation:
Yi, J. Experimental and Computational Approaches for Food Safety and Process Control. NC-1023 Multi-institutional Food Engineering Seminar, Mar 2022.
|