Cellular Actin Cytoskeleton Morphology Identification for Mechanical Characterization Using Deep Learning
Actin cytoskeleton morphology is able to affect and reflect the cellular mechanical properties. However, due to the lack of efficient approaches to quantifying actin cytoskeleton and mechanical properties, the study of mechanotransduction dynamics is still laborious. In this paper, a model to charac...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9874811/ |
| _version_ | 1828143520845660160 |
|---|---|
| author | Yi Liu Juntao Zhang Charuku Bharat Juan Ren |
| author_facet | Yi Liu Juntao Zhang Charuku Bharat Juan Ren |
| author_sort | Yi Liu |
| collection | DOAJ |
| description | Actin cytoskeleton morphology is able to affect and reflect the cellular mechanical properties. However, due to the lack of efficient approaches to quantifying actin cytoskeleton and mechanical properties, the study of mechanotransduction dynamics is still laborious. In this paper, a model to characterize the cellular actin cytoskeleton morphology was built using the graph to vector embedding technique together with neural network in machine learning (ML). The proposed ML model consists of a skip-gram model followed by a fully connected classifier. The images of NIH/3T3 cells treated with Latrunculin B at different concentrations were taken as the inputs, and the outputs were the actin cytoskeleton morphology labels defined by treatment concentrations (i.e., the actin depolymerization level). The proposed model was also compared to a general convolutional neural network (CNN) and three commonly used transfer learning models (GoogleNet, Xception, and VGG16), the results demonstrated the capabilities of the proposed model in extracting actin cytoskeleton features, avoiding overfitting, and keeping the model generalization. |
| first_indexed | 2024-04-11T19:58:34Z |
| format | Article |
| id | doaj.art-e283d465db06409c8eda14ab13e63fce |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-04-11T19:58:34Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-e283d465db06409c8eda14ab13e63fce2022-12-22T04:05:53ZengIEEEIEEE Access2169-35362022-01-0110974089741810.1109/ACCESS.2022.32037209874811Cellular Actin Cytoskeleton Morphology Identification for Mechanical Characterization Using Deep LearningYi Liu0Juntao Zhang1Charuku Bharat2Juan Ren3https://orcid.org/0000-0002-5616-7219Department of Mechanical Engineering, Iowa State University, Ames, IA, USADepartment of Mechanical Engineering, Iowa State University, Ames, IA, USADepartment of Mechanical Engineering, Iowa State University, Ames, IA, USADepartment of Mechanical Engineering, Iowa State University, Ames, IA, USAActin cytoskeleton morphology is able to affect and reflect the cellular mechanical properties. However, due to the lack of efficient approaches to quantifying actin cytoskeleton and mechanical properties, the study of mechanotransduction dynamics is still laborious. In this paper, a model to characterize the cellular actin cytoskeleton morphology was built using the graph to vector embedding technique together with neural network in machine learning (ML). The proposed ML model consists of a skip-gram model followed by a fully connected classifier. The images of NIH/3T3 cells treated with Latrunculin B at different concentrations were taken as the inputs, and the outputs were the actin cytoskeleton morphology labels defined by treatment concentrations (i.e., the actin depolymerization level). The proposed model was also compared to a general convolutional neural network (CNN) and three commonly used transfer learning models (GoogleNet, Xception, and VGG16), the results demonstrated the capabilities of the proposed model in extracting actin cytoskeleton features, avoiding overfitting, and keeping the model generalization.https://ieeexplore.ieee.org/document/9874811/Actin cytoskeletonclassificationgraph to vector embeddingmachine learningskip-gram model |
| spellingShingle | Yi Liu Juntao Zhang Charuku Bharat Juan Ren Cellular Actin Cytoskeleton Morphology Identification for Mechanical Characterization Using Deep Learning IEEE Access Actin cytoskeleton classification graph to vector embedding machine learning skip-gram model |
| title | Cellular Actin Cytoskeleton Morphology Identification for Mechanical Characterization Using Deep Learning |
| title_full | Cellular Actin Cytoskeleton Morphology Identification for Mechanical Characterization Using Deep Learning |
| title_fullStr | Cellular Actin Cytoskeleton Morphology Identification for Mechanical Characterization Using Deep Learning |
| title_full_unstemmed | Cellular Actin Cytoskeleton Morphology Identification for Mechanical Characterization Using Deep Learning |
| title_short | Cellular Actin Cytoskeleton Morphology Identification for Mechanical Characterization Using Deep Learning |
| title_sort | cellular actin cytoskeleton morphology identification for mechanical characterization using deep learning |
| topic | Actin cytoskeleton classification graph to vector embedding machine learning skip-gram model |
| url | https://ieeexplore.ieee.org/document/9874811/ |
| work_keys_str_mv | AT yiliu cellularactincytoskeletonmorphologyidentificationformechanicalcharacterizationusingdeeplearning AT juntaozhang cellularactincytoskeletonmorphologyidentificationformechanicalcharacterizationusingdeeplearning AT charukubharat cellularactincytoskeletonmorphologyidentificationformechanicalcharacterizationusingdeeplearning AT juanren cellularactincytoskeletonmorphologyidentificationformechanicalcharacterizationusingdeeplearning |