Dataset Information

Automated Deep Learning-Based System to Identify Endothelial Cells Derived from Induced Pluripotent Stem Cells.

ABSTRACT: Deep learning technology is rapidly advancing and is now used to solve complex problems. Here, we used deep learning in convolutional neural networks to establish an automated method to identify endothelial cells derived from induced pluripotent stem cells (iPSCs), without the need for immunostaining or lineage tracing. Networks were trained to predict whether phase-contrast images contain endothelial cells based on morphology only. Predictions were validated by comparison to immunofluorescence staining for CD31, a marker of endothelial cells. Method parameters were then automatically and iteratively optimized to increase prediction accuracy. We found that prediction accuracy was correlated with network depth and pixel size of images to be analyzed. Finally, K-fold cross-validation confirmed that optimized convolutional neural networks can identify endothelial cells with high performance, based only on morphology.

SUBMITTER: Kusumoto D

PROVIDER: S-EPMC5989816 | biostudies-literature | 2018 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Automated Deep Learning-Based System to Identify Endothelial Cells Derived from Induced Pluripotent Stem Cells.

Kusumoto Dai D Lachmann Mark M Kunihiro Takeshi T Yuasa Shinsuke S Kishino Yoshikazu Y Kimura Mai M Katsuki Toshiomi T Itoh Shogo S Seki Tomohisa T Fukuda Keiichi K

Stem cell reports 20180510 6

Deep learning technology is rapidly advancing and is now used to solve complex problems. Here, we used deep learning in convolutional neural networks to establish an automated method to identify endothelial cells derived from induced pluripotent stem cells (iPSCs), without the need for immunostaining or lineage tracing. Networks were trained to predict whether phase-contrast images contain endothelial cells based on morphology only. Predictions were validated by comparison to immunofluorescence ...[more]

PMID: 29754958

Similar Datasets

Project description:Endothelial cells derived from human induced pluripotent stem cells (hiPSC-ECs) provide a new opportunity for mechanistic research on vascular regeneration and drug screening. However, functions of hiPSC-ECs still need to be characterized. The objective of this study was to investigate electrophysiological and functional properties of hiPSC-ECs compared with primary human cardiac microvascular endothelial cells (HCMECs), mainly focusing on ion channels and membrane receptor signaling, as well as specific cell functions. HiPSC-ECs were derived from hiPS cells that were generated from human skin fibroblasts of three independent healthy donors. Phenotypic and functional comparison to HCMECs was performed by flow cytometry, immunofluorescence staining, quantitative reverse-transcription polymerase chain reaction (qPCR), enzyme-linked immunosorbent assay (ELISA), tube formation, LDL uptake, exosome release assays and, importantly, patch clamp techniques. HiPSC-ECs were successfully generated from hiPS cells and were identified by endothelial markers. The mRNA levels of KCNN2, KCNN4, KCNMA1, TRPV2, and SLC8A1 in hiPSC-ECs were significantly higher than HCMECs. AT1 receptor mRNA level in hiPSC-ECs was higher than in HCMECs. AT2 receptor mRNA level was the highest among all receptors. Adrenoceptor ADRA2 expression in hiPSC-ECs was lower than in HCMECs, while ADRA1, ADRB1, ADRB2, and G-protein GNA11 and Gai expression were similar in both cell types. The expression level of muscarinic and dopamine receptors CHRM3, DRD2, DRD3, and DRD4 in hiPSC-ECs were significantly lower than in HCMECs. The functional characteristics of endothelial cells, such as tube formation and LDL uptake assay, were not statistically different between hiPSC-ECs and HCMECs. Phenylephrine similarly increased the release of the vasoconstrictor endothelin-1 (ET-1) in hiPSC-ECs and HCMECs. Acetylcholine also similarly increased nitric oxide generation in hiPSC-ECs and HCMECs. The resting potentials (RPs), ISK1-3, ISK4 and IK1 were similar in hiPSC-ECs and HCMECs. IBK was larger and IKATP was smaller in hiPSC-ECs. In addition, we also noted a higher expression level of exosomes marker CD81 in hiPSC-ECs and a higher expression of CD9 and CD63 in HCMECs. However, the numbers of exosomes extracted from both types of cells did not differ significantly. The study demonstrates that hiPSC-ECs are similar to native endothelial cells in ion channel function and membrane receptor-coupled signaling and physiological cell functions, although some differences exist. This information may be helpful for research using hiPSC-ECs.

Dataset Information

Automated Deep Learning-Based System to Identify Endothelial Cells Derived from Induced Pluripotent Stem Cells.

Publications

Automated Deep Learning-Based System to Identify Endothelial Cells Derived from Induced Pluripotent Stem Cells.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets