Deep Learning Image Analysis of High-Throughput Toxicology Assay Images
High-throughput chemical screening approaches often employ microscopy to capture photomicrographs from multi-well cell culture plates, generating thousands of images that require time-consuming human analysis. To automate this subjective and time-consuming manual process, we have developed a method...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2022-01-01
|
| Series: | SLAS Discovery |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2472555221000149 |
| _version_ | 1818287114968629248 |
|---|---|
| author | Arpit Tandon Brian Howard Sreenivasa Ramaiahgari Adyasha Maharana Stephen Ferguson Ruchir Shah B. Alex Merrick |
| author_facet | Arpit Tandon Brian Howard Sreenivasa Ramaiahgari Adyasha Maharana Stephen Ferguson Ruchir Shah B. Alex Merrick |
| author_sort | Arpit Tandon |
| collection | DOAJ |
| description | High-throughput chemical screening approaches often employ microscopy to capture photomicrographs from multi-well cell culture plates, generating thousands of images that require time-consuming human analysis. To automate this subjective and time-consuming manual process, we have developed a method that uses deep learning to automatically classify digital assay images. We have trained a convolutional neural network (CNN) to perform binary and multi-class classification. The binary classifier binned assay images into healthy (comparable to untreated controls) and altered (not comparable to untreated-control) classes with >98% accuracy; the multi-class classifier assigned “Healthy,” “Intermediate” and “Altered” labels to assay images with >95% accuracy. Our dataset comprised high-resolution assay images from primary human hepatocytes and undifferentiated (proliferating) and differentiated 2D cultures of HepaRG cells. In this study we have focused on testing and fine-tuning various CNN architectures, including ResNet 34, 50 and 101. To visualize regions in the images that the CNN model used for classification, we employed Class Activation Maps (CAM). This allowed us to better understand the inner workings of the neural network and led to additional optimizations of the algorithm. The results indicate a strong correspondence between dosage and classifier-predicted scores, suggesting that these scores might be useful in further characterizing benchmark dose. Together, these results clearly demonstrate that deep-learning based automated image classification of cell morphology changes upon chemical-induced stress can yield highly accurate and reproducible assessments of cytotoxicity across a variety of cell types. |
| first_indexed | 2024-12-13T01:35:21Z |
| format | Article |
| id | doaj.art-8c6a90be8b5a4ef3855a69a9351185fd |
| institution | Directory Open Access Journal |
| issn | 2472-5552 |
| language | English |
| last_indexed | 2024-12-13T01:35:21Z |
| publishDate | 2022-01-01 |
| publisher | Elsevier |
| record_format | Article |
| series | SLAS Discovery |
| spelling | doaj.art-8c6a90be8b5a4ef3855a69a9351185fd2022-12-22T00:03:53ZengElsevierSLAS Discovery2472-55522022-01-012712938Deep Learning Image Analysis of High-Throughput Toxicology Assay ImagesArpit Tandon0Brian Howard1Sreenivasa Ramaiahgari2Adyasha Maharana3Stephen Ferguson4Ruchir Shah5B. Alex Merrick6Sciome LLC, Research Triangle Park, NC, USASciome LLC, Research Triangle Park, NC, USANational Toxicology Program Division, National Institute of Environmental Health Sciences, Research Triangle Park, NC, USASciome LLC, Research Triangle Park, NC, USANational Toxicology Program Division, National Institute of Environmental Health Sciences, Research Triangle Park, NC, USASciome LLC, Research Triangle Park, NC, USA; Correspondence should be addressed to.National Toxicology Program Division, National Institute of Environmental Health Sciences, Research Triangle Park, NC, USAHigh-throughput chemical screening approaches often employ microscopy to capture photomicrographs from multi-well cell culture plates, generating thousands of images that require time-consuming human analysis. To automate this subjective and time-consuming manual process, we have developed a method that uses deep learning to automatically classify digital assay images. We have trained a convolutional neural network (CNN) to perform binary and multi-class classification. The binary classifier binned assay images into healthy (comparable to untreated controls) and altered (not comparable to untreated-control) classes with >98% accuracy; the multi-class classifier assigned “Healthy,” “Intermediate” and “Altered” labels to assay images with >95% accuracy. Our dataset comprised high-resolution assay images from primary human hepatocytes and undifferentiated (proliferating) and differentiated 2D cultures of HepaRG cells. In this study we have focused on testing and fine-tuning various CNN architectures, including ResNet 34, 50 and 101. To visualize regions in the images that the CNN model used for classification, we employed Class Activation Maps (CAM). This allowed us to better understand the inner workings of the neural network and led to additional optimizations of the algorithm. The results indicate a strong correspondence between dosage and classifier-predicted scores, suggesting that these scores might be useful in further characterizing benchmark dose. Together, these results clearly demonstrate that deep-learning based automated image classification of cell morphology changes upon chemical-induced stress can yield highly accurate and reproducible assessments of cytotoxicity across a variety of cell types.http://www.sciencedirect.com/science/article/pii/S2472555221000149Deep learningToxicologyImage analysisCNN |
| spellingShingle | Arpit Tandon Brian Howard Sreenivasa Ramaiahgari Adyasha Maharana Stephen Ferguson Ruchir Shah B. Alex Merrick Deep Learning Image Analysis of High-Throughput Toxicology Assay Images SLAS Discovery Deep learning Toxicology Image analysis CNN |
| title | Deep Learning Image Analysis of High-Throughput Toxicology Assay Images |
| title_full | Deep Learning Image Analysis of High-Throughput Toxicology Assay Images |
| title_fullStr | Deep Learning Image Analysis of High-Throughput Toxicology Assay Images |
| title_full_unstemmed | Deep Learning Image Analysis of High-Throughput Toxicology Assay Images |
| title_short | Deep Learning Image Analysis of High-Throughput Toxicology Assay Images |
| title_sort | deep learning image analysis of high throughput toxicology assay images |
| topic | Deep learning Toxicology Image analysis CNN |
| url | http://www.sciencedirect.com/science/article/pii/S2472555221000149 |
| work_keys_str_mv | AT arpittandon deeplearningimageanalysisofhighthroughputtoxicologyassayimages AT brianhoward deeplearningimageanalysisofhighthroughputtoxicologyassayimages AT sreenivasaramaiahgari deeplearningimageanalysisofhighthroughputtoxicologyassayimages AT adyashamaharana deeplearningimageanalysisofhighthroughputtoxicologyassayimages AT stephenferguson deeplearningimageanalysisofhighthroughputtoxicologyassayimages AT ruchirshah deeplearningimageanalysisofhighthroughputtoxicologyassayimages AT balexmerrick deeplearningimageanalysisofhighthroughputtoxicologyassayimages |