Data from: High-content phenotyping reveals Golgi dynamics and their role in cell cycle regulation
Data files
Nov 07, 2025 version files 6.42 GB
-
cellphenotype_full.yaml
8.38 KB
-
Datasets.pdf
4.45 MB
-
Figure_1_S1.zip
707.81 MB
-
Figure_2BC_S2.zip
1.27 GB
-
Figure_2FGH_S3AB.zip
664.21 MB
-
Figure_3B_S3D.zip
268.90 MB
-
Figure_3CDE_S3CE.zip
1.09 GB
-
Figure_4A_S5A.zip
812.61 MB
-
Figure_4B.zip
452.76 MB
-
Figure_S4A.zip
637.40 MB
-
Figure_S4B.zip
97.78 MB
-
Figure_S4C.zip
401.26 MB
-
Figure_S5B.zip
7.51 MB
-
ptitprince_full.yaml
3.79 KB
-
README.md
7.66 KB
-
Readme.pdf
349.51 KB
-
Variables_TableS1.xlsx
11.68 KB
Abstract
Dataset DOI: 10.5061/dryad.8gtht771s
Description of the data and file structure
This dataset accompanies our study investigating the dynamic morphological changes of cellular structures, including the Golgi apparatus and cilia, during the cell cycle, using image-based high-content phenotyping. These files include the uncropped immunofluorescence (IF) image and Western blot (WB) data, cell profile data, and Python code, used for analysis of cell morphological features and figure generation, as described in Cao, Peng, Yang et al., J Cell Biol, 2025. The files also include the cell phenotypic profile datasets generated in this article.
For detailed instructions on running the code, reproducing analyses, and understanding the file structure with schematic diagrams, please refer to Readme.pdf.
The image data and code are organized according to the corresponding figures (e.g., Figure_1_S1.zip). Data and code include: image data (max intensity projection .tif files), segmentation results (.npy files), cell phenotype profiles (.csv files), and code (.ipynb Jupyter notebooks in the “Code” folder.
The code was executed in Miniconda using VS Code.
- To run code that includes raincloud plots, the
ptitprince_full.yamlAn environment is recommended. - Otherwise, all code should run with the
cellphenotype_full.yamlenvironment.
The cell phenotype features included in the cell profile CSV files are summarized in Table S1 of the article, and the source data file (Variables_TableS1.xlsx) for this table is also provided in this dataset for reference.
Files and variables
File: Figure_1_S1.zip
Description: Data and code for Figure 1 and Figure S1. HeLa cells were arrested at mitosis and treated with DMSO (control), taxol, or monastrol, followed by IF labeling for DNA, α-tubulin, and GM130 (Golgi). Segmentation results (.npy files) and cell phenotype profiles (.csv files) are included.
File: Figure_2BC_S2.zip
Description: Data and code for Figure 2B, C, and Figure S2. HeLa cells were arrested at G2 and treated with DMSO (control), SP, CytD, taxol, or BFA, followed by IF labeling for DNA, actin, Ki67 (proliferation marker), and GM130 (Golgi).
File: Figure_2FGH_S3AB.zip
Description: Data and code for Figure 2F–H and Figure S3A, B. ARPE-19 cells were serum-starved and stimulated, followed by IF labeling for DNA, actin, acetylated tubulin (cilia), and GRASP65 (Golgi).
File: Figure_3B_S3D.zip
Description: Data and code for Figure 3B and Figure S3D. Analysis of cell cycle phase (nuclear intensity of cell cycle markers) and ciliation rates in ARPE-19 cells transfected with siJNK1, siJNK2, or siAURKA, or treated with SP.
File: Figure_3CDE_S3CE.zip
Description: Data and code for Figure 3C–E and Figure S3, E. Similar experiments as Figure_2FGH_S3AB.zip, with siRNA transfection or drug treatment.
File: Figure_4A_S5A.zip
Description: Data and code for Figure 4A and Figure S5A. Quantification of centrosomal AURKA fluorescence signal in ARPE-19 cells under various conditions and time points.
File: Figure_4B.zip
Description: Data and code for Figure 4B. Measurement of spatial distances between the Golgi, nucleus, and centrosome in ARPE-19 cells.
File: Figure_S4A.zip
Description: Data and code for Figure S4A. siAURKA + siJNK2 double knockdown experiment.
File: Figure_S4B.zip
Description: Uncropped IF images for Figure S4B. Double labeling for IFT88 and acetylated tubulin (AcTub) to validate cilia detection.
File: Figure_S4C.zip
Description: Data and code for Figure S4C. Double labeling for GRASP65 and GM130 to validate Golgi detection and feature correlation. The analysis code and output data are included, which show a high correlation between the morphological features of the Golgi obtained from images of each marker.
File: Figure_S5B.zip
Description: WB images for Figure S5B, showing knockdown efficiency.
File: Datasets.pdf
Description: Datasets 1–7 in a single PDF. This file provides the phenotypic profiles obtained in this study. It is included here for reference and corresponds directly to the dataset descriptions and analyses presented in the article.
Dataset 1. Profiling results of metaphase-arrested cells for all analyzed features.
(A) Representative original images overlaid with color-coded labels assigned by DBSCAN. Label information is derived from Fig. 1F; note that the DMSO-treated cell image is identical to that in Fig. 1G.
(B) UMAP plots displaying the projection of values for all features used in the analysis. These UMAP plots are the same as those shown in Fig. 1E.
(C) Heatmap of the mean value of each feature for each cluster and experimental condition.
Dataset 2. Distribution profiles of analyzed features in metaphase-arrested cells via violin plots. Violin plots for all analyzed features are presented for the overall cell population as well as for each cluster.
Dataset 3. Cell phenotype profiles reflecting drug treatment effects on late G2 cells.
(A, B) All features used in the analysis are mapped onto the UMAP plot from Fig. 2B (A) and shown as violin plots by experimental condition (B). The complete dataset, including the features presented in.g S2E, F, is shown.
Dataset 4. Subcluster analysis of SP treatment effects on Golgi morphology in late G2 cells. Violin plots show the distributions of morphological features for two subclusters defined by Golgi morphology (see Fig. 2C) in both control and SP-treated groups. The complete dataset, including the features presented in Fig. 2C, is provided.
Dataset 5. Analysis of changes in cell phenotype profiles upon serum re-addition.
(A, B) All features used in the analysis are mapped onto the UMAP plot (A) and shown as violin plots grouped by clusters (B). Data from all time points are combined. Several key features highlighted in the text are marked with green boxes.
Dataset 6. Cell phenotype profiles reflecting the effects of kinase gene knockdown and SP treatment during the G0/G1 transition.
(A, B) All features used in the analysis are mapped onto the UMAP plot from Fig. 3D (A) and shown as violin plots grouped by experimental condition (B). The complete dataset, including the features presented in.g 3E, is provided.
Dataset 7. Cell phenotype profiles reflecting the effects of kinase gene knockdown and SP treatment during the G0/G1 transition, including siAURKA/siJNK2 double knockdown.
(A, B) All features used in the analysis are mapped onto the UMAP plot from Fig. S4A (A) and shown as violin plots grouped by experimental condition (B).
File: Readme.pdf
Description: Detailed instructions on data structure, analysis steps, and code execution.
File: cellphenotype_full.yaml
Description: A virtual environment to run all the code except the code containing raincloud plots.
File: ptitprince_full.yaml
Description: To run code that includes raincloud plots, this environment is recommended.
Variables: Variables_TableS1.xlsx
Description: This file provides the source data for Table S1 in the article. It lists and explains all variables (cell phenotypic features) that appear as column headers in the cell profile CSV files included in this dataset. Each variable, such as mask_area and mean_nuc, is accompanied by a brief description of its meaning, units and measurement.