Evaluation of Image Classification for Quantifying Mitochondrial Morphology Using Deep Learning.
Mitochondrial morphology
ResNet
deep learning
fission
fusion
mitochondrial dynamics
Journal
Endocrine, metabolic & immune disorders drug targets
ISSN: 2212-3873
Titre abrégé: Endocr Metab Immune Disord Drug Targets
Pays: United Arab Emirates
ID NLM: 101269157
Informations de publication
Date de publication:
2023
2023
Historique:
received:
13
12
2021
revised:
22
03
2022
accepted:
28
03
2022
pubmed:
6
7
2022
medline:
11
3
2023
entrez:
5
7
2022
Statut:
ppublish
Résumé
Mitochondrial morphology reversibly changes between fission and fusion. As these changes (mitochondrial dynamics) reflect the cellular condition, they are one of the simplest indicators of cell state and predictors of cell fate. However, it is currently difficult to classify them using a simple and objective method. The present study aimed to evaluate mitochondrial morphology using Deep Learning (DL) technique. Mitochondrial images stained by MitoTracker were acquired from HeLa and MC3T3-E1 cells using fluorescent microscopy and visually classified into four groups based on fission or fusion. The intra- and inter-rater reliabilities for visual classification were excellent [(ICC(1,3), 0.961 for rater 1; and 0.981 for rater 2) and ICC(1,3), respectively]. The images were divided into test and train images, and a 50-layer ResNet CNN architecture (ResNet-50) using MATLAB software was used to train the images. The datasets were trained five times based on five-fold cross-validation. The mean of the overall accuracy for classifying mitochondrial morphology was 0.73±0.10 in HeLa. For the classification of mixed images containing two types of cell lines, the overall accuracy using mixed images of both cell lines for training was higher (0.74±0.01) than that using different cell lines for training. We developed a classifier to categorize mitochondrial morphology using DL.
Sections du résumé
BACKGROUND
BACKGROUND
Mitochondrial morphology reversibly changes between fission and fusion. As these changes (mitochondrial dynamics) reflect the cellular condition, they are one of the simplest indicators of cell state and predictors of cell fate. However, it is currently difficult to classify them using a simple and objective method.
OBJECTIVE
OBJECTIVE
The present study aimed to evaluate mitochondrial morphology using Deep Learning (DL) technique.
METHODS
METHODS
Mitochondrial images stained by MitoTracker were acquired from HeLa and MC3T3-E1 cells using fluorescent microscopy and visually classified into four groups based on fission or fusion. The intra- and inter-rater reliabilities for visual classification were excellent [(ICC(1,3), 0.961 for rater 1; and 0.981 for rater 2) and ICC(1,3), respectively]. The images were divided into test and train images, and a 50-layer ResNet CNN architecture (ResNet-50) using MATLAB software was used to train the images. The datasets were trained five times based on five-fold cross-validation.
RESULT
RESULTS
The mean of the overall accuracy for classifying mitochondrial morphology was 0.73±0.10 in HeLa. For the classification of mixed images containing two types of cell lines, the overall accuracy using mixed images of both cell lines for training was higher (0.74±0.01) than that using different cell lines for training.
CONCLUSION
CONCLUSIONS
We developed a classifier to categorize mitochondrial morphology using DL.
Identifiants
pubmed: 35786342
pii: EMIDDT-EPUB-124938
doi: 10.2174/1871530322666220701093644
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
214-221Informations de copyright
Copyright© Bentham Science Publishers; For any queries, please email at epub@benthamscience.net.