Fig. 6

Quantitative comparison of 2.5D MGX and 2D SurfCut output. a–c Quantitative comparison of cell area on light-grown hypocotyl (same raw data as in Fig. 4b). a 2D analysis of cell area: cell contour extracted with SurfCut, cells segmented, and cell size quantified with PaCeQuant. b 2.5D analysis of cell area in the same light-grown hypocotyl using MGX. c Visual representation of the difference of cell size quantification between the two methods. Heatmaps in a and b represent the distribution of quantified cell size (μm²) while in c, it represents the difference in quantification between both methods. Positive values in c correspond to cases in which cell area quantified in 2.5D MGX is higher than in 2D SurfCut. d Scatter plot of the difference in cell size quantification relative to the angles of cell files in the tissue. The cells in the hypocotyl are organized in cell files. Each cell file is highlighted by a unique color and number corresponding to the colors and numbers in e. Here, each cell file was assigned a single average angle (see the “Methods” section). Cells “facing” the top view have a low angle. e–k Transverse optical section of the light-grown hypocotyl shown in a–c, showing in greyscale the raw confocal signal and in red the thin layer of signal extracted using SurfCut. f–l Close-ups (insets in e) highlighting one case in which the bias is almost completely absent (cell file 5 (f, i)) and two cases in which cell area measurement is biased (cell files 6 (g, j) and 8 (h, k)). The magenta and cyan lines in i–l are 1D representations of the MGX (magenta) and SurfCut/PaCeQuant (cyan) cell area quantification. l Schematic representation and explanation of the induced bias. As opposed to 2D SurfCut/PaCeQuant, cell area measurement in 2.5D with MGX fully accounts for the tissue- and cell-level curvature. Scale bars 50 μm