-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathexhibit.json
1 lines (1 loc) · 11.6 KB
/
exhibit.json
1
{"Images": [{"Name": "i0", "Description": "sample_1d.r1", "Path": "images/postMI-1d", "Width": 30238, "Height": 28381, "MaxLevel": 5}], "Header": "This sample depicts a cross section of a murine heart acquired 1 day post myocardial infarction induction. Imaging was performed using the Lunaphore COMET (PA) platform with a 16plex antibody panel.\n\nRaw data available at: https://www.synapse.org/Synapse:syn51449054/files/\nManuscript: https://doi.org/10.1101/2024.05.20.594955", "Rotation": 0, "Layout": {"Grid": [["i0"]]}, "Stories": [{"Name": "", "Description": "", "Waypoints": [{"Name": "1 day post-MI - Overview", "Description": "In the acute phase of myocardial infarction (MI) progression, the immune response and infiltration play a key role in cardiac healing and subsequent heart functionality. Utilizing a minimally invasive MI induction model and combining targeted transcriptomic and proteomic approaches on heart cross sections in the acute phase of MI, we were able to observe a novel mode of immune infiltration into the injury region - through the endocardial layer.\n\nOur findings report that trans-endocardial infiltration is most prominent at 24 hours post MI induction which is why we present the Minerva story for this time point.", "Arrows": [], "Overlays": [], "Group": "Overview", "Masks": [], "ActiveMasks": [], "Zoom": 0.5813265306122446, "Pan": [0.3049390036106875, 0.5400761464704668]}, {"Name": "Region annotation", "Description": "Based on the cardiac geometry, Tnnt2 and Ankrd1 signal presence, and proximity to the endocardial layer we've annotated the following regions:\n- LV lumen: Left ventricular lumen.\n- Infarct core: annotated as the region characterized by low Tnnt2 and high Ankrd1 intensity in which the infarct is transmural, additionally guided by heart geometry if the infarct is not transmural.\n- Border zone: Encapsulates non-infarct core injured regions, as well as the surrounding 5-cell thick layer of uninjured cells.\n- Epicardial region: adjacent to the infarct core, the 3 to 5 cell thick region adjacent to the epicardial layer.\n- Near-MI endocardial region: adjacent to the infarct core, the region between the lumen and injured tissue. This region includes the endocardial layer, as well as adjacent uninjured Tnnt2+ cardiomyocytes.\n- Remote endocardial region: near-endocard region in an area remote of the infarct annotated with a comparable thickness to the near-MI endocardial region.\n- Background: Imaged area not containing cardiac tissue.\n- Artefact: area with artefacts - folds, dirt, bubbles, out-of-focus regions.\n\n\n", "Arrows": [{"Text": "LV lumen", "HideArrow": false, "Point": [0.5426720393405504, 0.6366800503836265], "Angle": 183}, {"Text": "Infarct core", "HideArrow": false, "Point": [0.5567145178380053, 0.8762798397464492], "Angle": 153}, {"Text": "Epicardial region", "HideArrow": false, "Point": [0.8215419704375813, 0.6705950122488985], "Angle": 60}, {"Text": "Border zone", "HideArrow": false, "Point": [0.6117920390413142, 0.29507592656845316], "Angle": 60}, {"Text": "Remote endocardial region", "HideArrow": false, "Point": [0.35316806185041233, 0.4945033377895969], "Angle": 225}, {"Text": "Near-MI endocardial region", "HideArrow": false, "Point": [0.6327756576608058, 0.751718833482097], "Angle": 273}, {"Text": "Artefact", "HideArrow": false, "Point": [0.6373202545947747, 0.8808898372406009], "Angle": 35}, {"Text": "Background", "HideArrow": false, "Point": [0.7784589228404437, 0.1707403257764324], "Angle": 207}], "Overlays": [], "Group": "Cardiomyocytes", "Masks": ["All regions"], "ActiveMasks": ["All regions"], "Zoom": 0.3841537424140256, "Pan": [0.5327155491349846, 0.5253228344650867]}, {"Name": "Segmentation", "Description": "With the diverse tissue landscape in the context of cell types, shapes and characteristics, segmentation is challenging.\n\nOnly utilizing the nuclear staining with DAPI for cell segmentation, would result in most of the area not being captured, as a high percentage of the tissue consists of cardiomyocytes, the cardiac muscle cells. These cells can be very large and polynucleated. Depending on their position in the cross section, it is possible to not capture a nucleus within the area denoted by their membrane.\n\nWe have used WGA (wheat-germ-agglutinin) as a membrane marker. To counteract the fact it stains fibrotic regions strongly, and to aid with segmentation, we've applied contrast-limited adaptive histogram equalization (CLAHE) prior to segmentation.\n\nExisting off-the-shelf state-of-the-art models did not perform well on our data which is why we have interactively trained a Cellpose 2 model on contast-adjusted DAPI and WGA images of our data that could capture cardiomyocytes, as well as smaller cells well-characterized by their nucleus.", "Arrows": [], "Overlays": [], "Group": "Segmentation", "Masks": [], "ActiveMasks": [], "Zoom": 7.102397996797181, "Pan": [0.5609440958866988, 0.46859635014936785], "VisMatrix": {"config": {"view": {"continuousWidth": 400, "continuousHeight": 300}, "background": null}, "data": {"url": "data/minerva_segmentation.csv"}, "mark": "rect", "encoding": {"color": {"field": "frequency", "type": "quantitative", "legend": {"direction": "horizontal", "orient": "bottom"}}, "x": {"field": "channel", "sort": ["meanIOU", "meanDice", "Precision", "Recall", "F1"], "type": "nominal", "grid": false}, "y": {"field": "type", "sort": ["Mesmer nuclear", "Mesmer wholecell", "Cellpose cyto", "Cellpose custom"], "type": "nominal", "grid": false}}, "$schema": "https://vega.github.io/schema/vega-lite/v4.17.0.json", "datasets": {}, "width": "container"}}, {"Name": "Immune landscape", "Description": "The immune landscape of the acute MI is diverse and dynamic. Consisting of resident macrophages, infiltrating monocytes/macrophages, neutrophils and other myeloid cells with their specific functions, detailed characterization can be difficult.\n\nNeutrophils are known to be the first immune cells to infiltrate, closely followed by Ccr2+ monocytes/macrophages.\n\nIt had been previously shown that immune cells infiltrate through the epicardial layer from the pericardium and from the border zones through vasculature.\n\nHere we detect endocardial layer-specific infiltration of immune cells entering the heart from the inside of the LV lumen.", "Arrows": [{"Text": "Border zone infiltration", "HideArrow": false, "Point": [0.33251449257091625, 0.741347838586199], "Angle": 147}, {"Text": "Endocardial infiltration", "HideArrow": false, "Point": [0.45160503678929353, 0.8288081780171996], "Angle": 60}, {"Text": "Epicardial infiltration", "HideArrow": false, "Point": [0.36266507016393473, 0.8711659546318643], "Angle": 191}], "Overlays": [], "Group": "All_immune", "Masks": [], "ActiveMasks": [], "Zoom": 1.6544793090104795, "Pan": [0.34163494732001354, 0.7993439773731704]}, {"Name": "Immune infiltration through the near-infarct endocardium", "Description": "Here we see a region of the near-infarct endocardium.\n\nWith an endothelial marker (Cd31), we can see clearly defined endothelial cells making up the microvasculature. Cd31 also stains the endocardial layer because endocardial cells are also endothelial cells.\n\nWhat can be observed is an infiltration of immune cells, especially Ccr2+ Monocytes/Macrophages through the endocardial layer. Taking the proximity of microvasculature endothelial cells, it is impossible for these cells to have infiltrated through the microvasculature, especially knowing the direction of movement is towards the infarct.", "Arrows": [{"Text": "", "HideArrow": false, "Point": [0.4492982960484438, 0.8288414190203597], "Angle": 60}, {"Text": "", "HideArrow": false, "Point": [0.589991377077052, 0.8375491833230385], "Angle": 60}, {"Text": "", "HideArrow": false, "Point": [0.5377693904954892, 0.8435689078968204], "Angle": 60}, {"Text": "", "HideArrow": false, "Point": [0.4782031061738144, 0.8208884280997315], "Angle": 60}], "Overlays": [{"x": 0.42729758182902766, "y": 0.7727070935752022, "width": 0.20811277112635407, "height": 0.08134357044024998}], "Group": "Ccr2 Mpo Cd68 and Cd31", "Masks": [], "ActiveMasks": [], "Zoom": 4.1235747713106, "Pan": [0.5290026322689342, 0.8068595635200217]}, {"Name": "No immune infiltration through the remote-infarct endocardium", "Description": "When observing a comparable region remote to the infarct, infiltration events through the endocardial layer are much less pronounced.", "Arrows": [], "Overlays": [{"x": 0.33955956095310214, "y": 0.4288189309016104, "width": 0.06428506897667541, "height": 0.1607833153746301}], "Group": "Ccr2 Mpo Cd68 and Cd31", "Masks": [], "ActiveMasks": [], "Zoom": 5.201038554131691, "Pan": [0.3765761772198523, 0.4785703336022066]}, {"Name": "Additional infiltration points", "Description": "", "Arrows": [], "Overlays": [{"x": 0.6543983795424778, "y": 0.6698121525484502, "width": 0.054452530412838485, "height": 0.0583960172383714}], "Group": "Ccr2 Mpo Cd68 and Cd31", "Masks": [], "ActiveMasks": [], "Zoom": 6.241246264958032, "Pan": [0.6709826323567054, 0.6987400593302706]}, {"Name": "Additional infiltration points", "Description": "", "Arrows": [], "Overlays": [{"x": 0.6746019969774001, "y": 0.5459434499327376, "width": 0.06298774847358102, "height": 0.06720133713647936}], "Group": "Ccr2 Mpo Cd68 and Cd31", "Masks": [], "ActiveMasks": [], "Zoom": 6.241246264958032, "Pan": [0.6859462741467415, 0.5793010268473485]}, {"Name": "Additional infiltration points", "Description": "", "Arrows": [], "Overlays": [{"x": 0.63296503419492, "y": 0.5024702724931448, "width": 0.093365201782595, "height": 0.03916476641796418}], "Group": "Ccr2 Mpo Cd68 and Cd31", "Masks": [], "ActiveMasks": [], "Zoom": 7.48949551794964, "Pan": [0.6834740777822255, 0.5157727879833843]}]}], "Channels": [{"Rendered": false, "Name": "DAPI", "Path": "DAPI_0__DAPI"}, {"Rendered": false, "Name": "TNNT2_3000_Cy5", "Path": "TNNT2-3000-Cy5_16__TNNT2-3000-Cy5"}, {"Rendered": false, "Name": "ANKRD1_200_Cy5", "Path": "ANKRD1-200-Cy5_15__ANKRD1-200-Cy5"}, {"Rendered": false, "Name": "CD31_250_Cy5", "Path": "CD31-250-Cy5_5__CD31-250-Cy5"}, {"Rendered": false, "Name": "aSMA_400_Cy5", "Path": "aSMA-400-Cy5_3__aSMA-400-Cy5"}, {"Rendered": false, "Name": "TREM_150_Cy5", "Path": "TREM-150-Cy5_1__TREM-150-Cy5"}, {"Rendered": false, "Name": "CD68_300_TRITC", "Path": "CD68-300-TRITC_2__CD68-300-TRITC"}, {"Rendered": false, "Name": "CCR2_250_Cy5", "Path": "CCR2-250-Cy5_7__CCR2-250-Cy5"}, {"Rendered": false, "Name": "MPO_150_Cy5", "Path": "MPO-150-Cy5_13__MPO-150-Cy5"}, {"Rendered": false, "Name": "CD45_200_TRITC", "Path": "CD45-200-TRITC_6__CD45-200-TRITC"}, {"Rendered": false, "Name": "WGA_250_TRITC", "Path": "WGA-250-TRITC_17__WGA-250-TRITC"}], "Masks": [{"Path": "All-regions", "Name": "All regions", "Colors": ["ffffff"], "Channels": ["All regions"]}], "Groups": [{"Name": "Overview", "Colors": ["00ffff", "0000ff", "ff6600", "ffff00", "ff0000"], "Channels": ["DAPI", "TNNT2_3000_Cy5", "ANKRD1_200_Cy5", "CD31_250_Cy5", "aSMA_400_Cy5"], "Descriptions": ["", "", "", "", ""]}, {"Name": "All_immune", "Colors": ["00ffff", "00ff00", "ff00ec", "ff0000", "0800ff"], "Channels": ["TREM_150_Cy5", "CD68_300_TRITC", "CCR2_250_Cy5", "MPO_150_Cy5", "CD45_200_TRITC"], "Descriptions": ["", "", "", "", ""]}, {"Name": "Ccr2 Mpo Cd68 and Cd31", "Colors": ["ff00ec", "ff0000", "00ff00", "ffff00"], "Channels": ["CCR2_250_Cy5", "MPO_150_Cy5", "CD68_300_TRITC", "CD31_250_Cy5"], "Descriptions": ["", "", "", ""]}, {"Name": "Cardiomyocytes", "Colors": ["0000ff", "ff6600"], "Channels": ["TNNT2_3000_Cy5", "ANKRD1_200_Cy5"], "Descriptions": ["", ""]}, {"Name": "Segmentation", "Colors": ["00ffff", "ff00ff"], "Channels": ["DAPI", "WGA_250_TRITC"], "Descriptions": ["", ""]}], "PixelsPerMicron": 4.3478260869565215, "FirstGroup": "Overview", "FirstViewport": {"Pan": [0.2775812679343528, 0.5007898894154819], "Zoom": 0.5382653061224489}}