Vinc/E-cad ratio correlates with the junctional tension. a Representative image used as input for segmentation (scale bar 5 μm). b The “skeleton” image with the approximated network of junctions used as input for mechanical inference (see Methods). c The output of mechanical inference where line thickness indicates junctional tension. d The planar polarized distribution of inferred tension. Number of embryos in inset. e, f Scatter plot showing the distribution of the pre-ablation Vinc/E-cad ratio against e the post-ablation recoil velocities for multiple events of junctional ablations and f the pre-ablation inferred tension, for a subset of junctions in e where mechanical inference was feasible. In insets, Spearman correlation coefficient ρ, corresponding p-value, number of ablations events n, pooled from N embryos. g Scatter plot showing the distribution of Vinc/E-cad ratio against inferred tension for junctions from a single image. In inset, Spearman correlation coefficient ρ, corresponding p-value, and number of junctions n. h Box plots showing the distribution of conditional correlation coefficients for Vinc/E-cad ratio, Vinculin and junctional Myosin-II intensity against inferred tension conditioned on the length of the junction. i Same data as h, plotted as a function of junction length (also, see Methods). The statistics are based on 37,350 junctions across six embryos. j–l Box plots showing the distribution of local correlation coefficients (see Methods) between inferred tension and either the Junctional Myosin-II (j, n = 3500 cells), the Vinc/E-cad ratio (k, n = 9000 cells) or the Vinculin intensity (l, n = 9000 cells), for WT embryos, H1152-injected embryos along with statistical null. Vinculin density and Vinc/E-cad ratio related quantifications come from embryos co-expressing Vinculin-mCherry and E-cadherin-GFP (n = 6 embryos for WT, as well as Rok inhibition), while those for junctional Myosin-II come from embryos co-expression MyoII-mCherry and E-cadherin-GFP (n = 5 embryos for WT and n = 4 embryos for Rok inhibition). The error bars in d represent SEM. The boxes in h and j–l represent 25th to 75th percentiles; the whiskers represent 5th to 95th percentiles; the red lines represent the medians; and p-values estimated using Mann–Whitney U-test. The error bars in i represent the standard error across 100 different bins with the same length of junction