Utilization of convolutional neural networks to analyze microscopic images for high-throughput screening of mesenchymal stem cells

IF 1.7 4区 生物学 Q3 BIOLOGY Open Life Sciences Pub Date : 2024-07-10 DOI:10.1515/biol-2022-0859
MuYun Liu, XiangXi Du, JunYuan Hu, Xiao Liang, HaiJun Wang
{"title":"Utilization of convolutional neural networks to analyze microscopic images for high-throughput screening of mesenchymal stem cells","authors":"MuYun Liu, XiangXi Du, JunYuan Hu, Xiao Liang, HaiJun Wang","doi":"10.1515/biol-2022-0859","DOIUrl":null,"url":null,"abstract":"This work investigated the high-throughput classification performance of microscopic images of mesenchymal stem cells (MSCs) using a hyperspectral imaging-based separable convolutional neural network (CNN) (H-SCNN) model. Human bone marrow mesenchymal stem cells (hBMSCs) were cultured, and microscopic images were acquired using a fully automated microscope. Flow cytometry (FCT) was employed for functional classification. Subsequently, the H-SCNN model was established. The hyperspectral microscopic (HSM) images were created, and the spatial-spectral combined distance (SSCD) was employed to derive the spatial-spectral neighbors (SSNs) for each pixel in the training set to determine the optimal parameters. Then, a separable CNN (SCNN) was adopted instead of the classic convolutional layer. Additionally, cultured cells were seeded into 96-well plates, and high-functioning hBMSCs were screened using both manual visual inspection (MV group) and the H-SCNN model (H-SCNN group), with each group consisting of 96 samples. FCT served as the benchmark to compare the area under the curve (AUC), <jats:italic>F</jats:italic>1 score, accuracy (Acc), sensitivity (Sen), specificity (Spe), positive predictive value (PPV), and negative predictive value (NPV) between the manual and model groups. The best classification Acc was 0.862 when using window size of 9 and 12 SSNs. The classification Acc of the SCNN model, ResNet model, and VGGNet model gradually increased with the increase in sample size, reaching 89.56 ± 3.09, 80.61 ± 2.83, and 80.06 ± 3.01%, respectively at the sample size of 100. The corresponding training time for the SCNN model was significantly shorter at 21.32 ± 1.09 min compared to ResNet (36.09 ± 3.11 min) and VGGNet models (34.73 ± 3.72 min) (<jats:italic>P</jats:italic> &lt; 0.05). Furthermore, the classification AUC, <jats:italic>F</jats:italic>1 score, Acc, Sen, Spe, PPV, and NPV were all higher in the H-SCNN group, with significantly less time required (<jats:italic>P</jats:italic> &lt; 0.05). Microscopic images based on the H-SCNN model proved to be effective for the classification assessment of hBMSCs, demonstrating excellent performance in classification Acc and efficiency, enabling its potential to be a powerful tool in future MSCs research.","PeriodicalId":19605,"journal":{"name":"Open Life Sciences","volume":"53 1","pages":""},"PeriodicalIF":1.7000,"publicationDate":"2024-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Open Life Sciences","FirstCategoryId":"99","ListUrlMain":"https://doi.org/10.1515/biol-2022-0859","RegionNum":4,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"BIOLOGY","Score":null,"Total":0}
引用次数: 0

Abstract

This work investigated the high-throughput classification performance of microscopic images of mesenchymal stem cells (MSCs) using a hyperspectral imaging-based separable convolutional neural network (CNN) (H-SCNN) model. Human bone marrow mesenchymal stem cells (hBMSCs) were cultured, and microscopic images were acquired using a fully automated microscope. Flow cytometry (FCT) was employed for functional classification. Subsequently, the H-SCNN model was established. The hyperspectral microscopic (HSM) images were created, and the spatial-spectral combined distance (SSCD) was employed to derive the spatial-spectral neighbors (SSNs) for each pixel in the training set to determine the optimal parameters. Then, a separable CNN (SCNN) was adopted instead of the classic convolutional layer. Additionally, cultured cells were seeded into 96-well plates, and high-functioning hBMSCs were screened using both manual visual inspection (MV group) and the H-SCNN model (H-SCNN group), with each group consisting of 96 samples. FCT served as the benchmark to compare the area under the curve (AUC), F1 score, accuracy (Acc), sensitivity (Sen), specificity (Spe), positive predictive value (PPV), and negative predictive value (NPV) between the manual and model groups. The best classification Acc was 0.862 when using window size of 9 and 12 SSNs. The classification Acc of the SCNN model, ResNet model, and VGGNet model gradually increased with the increase in sample size, reaching 89.56 ± 3.09, 80.61 ± 2.83, and 80.06 ± 3.01%, respectively at the sample size of 100. The corresponding training time for the SCNN model was significantly shorter at 21.32 ± 1.09 min compared to ResNet (36.09 ± 3.11 min) and VGGNet models (34.73 ± 3.72 min) (P < 0.05). Furthermore, the classification AUC, F1 score, Acc, Sen, Spe, PPV, and NPV were all higher in the H-SCNN group, with significantly less time required (P < 0.05). Microscopic images based on the H-SCNN model proved to be effective for the classification assessment of hBMSCs, demonstrating excellent performance in classification Acc and efficiency, enabling its potential to be a powerful tool in future MSCs research.
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
利用卷积神经网络来分析显微图像,从而实现间充质干细胞的高通量筛选
这项研究利用基于高光谱成像的可分离卷积神经网络(CNN)(H-SCNN)模型,研究了间充质干细胞(MSCs)显微图像的高通量分类性能。培养人骨髓间充质干细胞(hBMSCs),并使用全自动显微镜获取显微图像。采用流式细胞术(FCT)进行功能分类。随后,建立了 H-SCNN 模型。创建了高光谱显微(HSM)图像,并利用空间-光谱组合距离(SSCD)得出了训练集中每个像素的空间-光谱邻域(SSN),从而确定了最佳参数。然后,采用可分离 CNN(SCNN)代替传统的卷积层。此外,将培养细胞播种到 96 孔板中,使用人工目测(MV 组)和 H-SCNN 模型(H-SCNN 组)筛选高功能 hBMSCs,每组包括 96 个样本。以 FCT 为基准,比较人工组和模型组的曲线下面积(AUC)、F1 分数、准确度(Acc)、灵敏度(Sen)、特异度(Spe)、阳性预测值(PPV)和阴性预测值(NPV)。当使用 9 个和 12 个 SSN 的窗口大小时,最佳分类 Acc 为 0.862。随着样本量的增加,SCNN 模型、ResNet 模型和 VGGNet 模型的分类加速度逐渐增加,在样本量为 100 时分别达到 89.56 ± 3.09%、80.61 ± 2.83% 和 80.06 ± 3.01%。与 ResNet 模型(36.09 ± 3.11 分钟)和 VGGNet 模型(34.73 ± 3.72 分钟)相比,SCNN 模型的相应训练时间明显缩短,为 21.32 ± 1.09 分钟(P < 0.05)。此外,H-SCNN 组的分类 AUC、F1 分数、Acc、Sen、Spe、PPV 和 NPV 都更高,所需的时间也明显更少(P <0.05)。事实证明,基于 H-SCNN 模型的显微图像能有效地对 hBMSCs 进行分类评估,在分类速度和效率方面表现出色,有望成为未来间充质干细胞研究的有力工具。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
CiteScore
2.50
自引率
4.50%
发文量
131
审稿时长
43 weeks
期刊介绍: Open Life Sciences (previously Central European Journal of Biology) is a fast growing peer-reviewed journal, devoted to scholarly research in all areas of life sciences, such as molecular biology, plant science, biotechnology, cell biology, biochemistry, biophysics, microbiology and virology, ecology, differentiation and development, genetics and many others. Open Life Sciences assures top quality of published data through critical peer review and editorial involvement throughout the whole publication process. Thanks to the Open Access model of publishing, it also offers unrestricted access to published articles for all users.
期刊最新文献
Preliminary analysis of the role of small hepatitis B surface proteins mutations in the pathogenesis of occult hepatitis B infection via the endoplasmic reticulum stress-induced UPR-ERAD pathway. Optimizing a modified cetyltrimethylammonium bromide protocol for fungal DNA extraction: Insights from multilocus gene amplification. Proteome differences of dental stem cells between permanent and deciduous teeth by data-independent acquisition proteomics. Comparative analysis of inflammatory biomarkers for the diagnosis of neonatal sepsis: IL-6, IL-8, SAA, CRP, and PCT. Post-pandemic insights on COVID-19 and premature ovarian insufficiency.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1