An in vivo microscopy dataset capturing leukocyte cell death

Cell death is pivotal for multiple biological processes including tissue homeostasis, embryological development, and host defense. Recent advancements in intravital microscopy have enabled the study of cell death in vivo under a variety of experimental conditions, such as infection and cancer. However, the limited throughput of this technology, together with a lack of openly accessible datasets, circumscribes the extent of cell death research in vivo. This problem affects the development of algorithms for the automatic detection and characterization of cell death, which requires the integration of extensive and curated datasets. To address these needs, we present a dataset of immune cells undergoing cell death carefully selected and annotated for the tasks of classification, detection, and segmentation. The dataset encompasses neutrophils, eosinophils, and dendritic cells acquired within the spleen and lymph node of mice. Moreover, the dataset provides spatial coordinates, temporal annotation, and 3D reconstruction of each cell death.

Description of the data and file structure

The presented dataset contains 4D time-lapses (x, y, z, t, c) generated using two-photon intravital microscopy (2P-IVM). The videos included in the dataset consist of unprocessed raw microscopy footage, each one comprehensive of a unique identifier (ID) and essential metadata information. Each video spans five dimensions: spatial coordinates (x, y, z), time (t), and acquisition channels (c). Raw videos are available in HDF5 format (Imaris) and divided according to the cell type in three different folders:

Dendritics_movies

Eosinophils_movies

Neutrophils_movies

Each HDF5 file contains two types of annotations in the form of 1) tracks coordinates of cell death events (x, y, z, t, ID), and 2) 3D volumetric reconstructions of dying cells (meshes described by vertices and edges).

An additional folder entitled "Annotations_CSV.zip" contains two sub-folders:

Cell deaths

Semantic annotations

"Cell deaths" contains the trajectories of death events in Excel format (CSV). Each filename contains the identifier of the corresponding microscopy acquisition, and the channel in which the events were observed, e.g. "Neu1_ch1".

The folder "Semantic annotations" contains the semantic labels (CSV) describing the morphological state of cells at different stages of the cell death process. This notation indicates if dying cells have a "blebbing" or "disruptive" morphology and if they elicit an immune response. Start and ending frames are reported to provide the time interval duration of each state.

Finally, three additional folders contain crops of microscopy movies centered around individual deaths. These are found in:

Dendritics_crops.zip

Eosinophils_crops.zip

Neutrophil_crops.zip

Each cropped movie is named after the original uncropped acquisition, and it specifies the cell ID of the dying cell depicted, e.g. "Neu1_cell1".

Sharing/Access information

Data are also available at: www.immunemap.org

Code to process the dataset and demo available at: https://github.com/AlainPulfer/Leuko-Death.git