Figure 3 - computationally advected segmentation, for measuring influence of T1 transitions on constrictions

This dataset is the initial segmentation measured at the onset of the first constriction to appear (the middle constriction in space) that is then advected according to the meso-scale (tissue-scale) cellular flow. No topological rearrangements are allowed in the computational advection. The utility of this approach is to compare the advected cell shapes computed here against the true cell shapes in order to infer the presence and orientation of T1 transitions on morphogenesis in the absence of cell divisions and extrusions. Here, the initial segmentation spanning most of the endoderm of the midgut is advected in the mean flow. <br> The dataset is structured as follows: For a given semented cell (cellID) segmented at timepoint t=0 to a different timepoint t, segVertexPathlines2D(timePoints==t, cellIDs==cellID, 1) is the 2D pullback position in the s direction and segVertexPathlines2D(timePoints=t, cellIDs==cellID, 2) is the 2D pullback position in the phi direction. segVertexPathlines3D is the corresponding embedding space (lab space / embryo space) positions of those cells. <br> In more detail, timePoints is a list of timestamps in minutes relative to the onset of the first constriction. cellIDs are a list of indices for segVertexPathlines2D/3D to indicate which cell a given set of segmentation points corresponds to. The true 3D segmentation is then given by segVertexPathlines3D, with each cell being a subset of the array. The first index indexes time, the second runs through all cells and can be filtered for a given cell using cellIDs==cellID_desired, and the last index is over dimensions of space.