semi-automated image and data analysis workflows  (Sensovation ag)

Journal: Cerebellum (London, England)

Article Title: Innovative 3D-Image Analysis of Cerebellar Vascularization Highlights Angiogenic Gene Dysregulations in a Murine Model of Apnea of Prematurity

doi: 10.1007/s12311-026-02006-1

Figure Lengend Snippet: Images illustrating the Imaris workflow developed for the vascular network modeling. A–I: Sequential workflow steps allowing the analysis of the cerebellar vascular network of a P4 mouse cerebellum on the Imaris software. From a 3D lightsheet acquisition (A) , the cerebellum is delineated (B) and a mask is created (C) . Within that selected volume, the cerebellar vascularization is segmented (D) , which allows the network visualization (E) and the separation of a deep and a superficial network (F) . Then the threshold of seedpoints is defined (G) , and thanks to the artificial intelligence module (AI), Imaris is able to discriminate “true” (blue) and “false” (red) seedpoints (H) , and “true” (blue) and “false” (red) segments (I) . AI: artificial intelligence; Px: postnatal day x

Article Snippet: Moreover, to correlate putative modifications of gene expression with defects in vessel morphology, we developed an image analysis workflow using both Imaris and VesselVio software to visualize the cerebellar vascularization in 3D during postnatal development and obtain comparable quantitative parameters between normoxic and IH conditions.

Techniques: Software

Journal: bioRxiv

Article Title: Unraveling the Complexity of Abdominal Aortic Aneurysm: Multiplexed Imaging Insights into C-Reactive Protein-Related Variations

doi: 10.1101/2024.02.22.581315

Figure Lengend Snippet: (A) Schematic overview of the CODEX imaging workflow. We constructed 31 antibody panels tagged by oligonucleotides. All antibodies were stained in a single step, and imaging was performed on every three antibodies in a cyclic reveal-imaging-remove manner. All images were preprocessed and layered in one qptiff multidimensional image. (B) Computational analysis pipeline for multiplexed imaging data. After CODEX imaging, we segmented DAPI nuclei using Mesmer’s unsupervised segmentation method. Then, we extracted the mean intensity from each segmented cell, making 415,365 cells × 31 Antibodies matrix. Scatter plots for each antibody pair were visualized for quality control purposes. The area was calculated by binarized image for each ROI. Unsupervised clustering was performed for 415,365 cells using z score normalized mean intensity of 29 antibodies. We used the GPU version of Phenograph, Grapheno, for the unsupervised clustering. Fifty-one distinct phenotypes were identified. The right column contains narrative annotations for each clustering result. The heatmap was visualized by z expression score. Antibodies were also clustered, and same lineage antibodies were also clustered, which shows biological relevance. The left bar graph shows each cluster’s cell composition (%) of three groups: High-CRP, Low-CRP, and normal control. The composition ratio varies according to the phenotype cluster.

Article Snippet: The overview of the CODEX multiplexed imaging computational analysis workflow is illustrated in .

Techniques: Imaging, Construct, Staining, Expressing

Journal: Frontiers in Neuroscience

Article Title: Advancements in high-resolution 3D microscopy analysis of endosomal morphology in postmortem Alzheimer’s disease brains

doi: 10.3389/fnins.2023.1321680

Figure Lengend Snippet: Overview of Imaris image analysis pipeline: (A) First, parameters are set for neuronal structure identification and segmentation via four separate “object creation” workflows using several representative z-stacks from the AD case series. Two Surface object workflows are created to define the neuronal regions-of-interest (ROIs) in which early endosomes were observed: the MAP2+ somatodendritic neuronal ROI, and neuronal DAPI+ perinuclear ROI. Both Surface object and Spots object workflows are used to gather quantitative data on EEA1+ early endosome puncta, as the Surface and Spots object workflows provide unique statistical variables related to endosome size, morphology, density, and spatial distribution within the neuronal ROIs. The workflows in (A) are then combined into batch pipelines, as shown in (B) to allow for analysis of the entire AD case cohort. First the neuronal ROIs are created using Batch Pipeline 1 (B1), after which the ROIs are refined to exclude disconnected neurites in the neuropil surrounding the neuron imaged from the MAP2+ surface and to remove non-neuronal nuclei from the DAPI+ surface (B2). The EEA1+ green channel is then masked twice over, using the two neuronal ROIs created in steps (B1-B2). Lastly, the endosome analysis batch pipeline (Batch pipeline 2) can be built, comprised of the EEA1+ surface and spots workflows created in (A) , with both workflows duplicated to analyze both the masked EEA1 channels (B4).

Article Snippet: Two separate image analysis workflows offered by Imaris are used for EEA1(+) early endosome analysis: the 3D “surface” and “spots” object rendering tools ( , ).

Techniques: