Figure 5.

Inclusion of an exponentially graded distribution in cortex membrane linker strength greatly improves model performance for elongated cells chemotaxing in low resistive environments (0.7% agarose). (a,b) Frequency of observed blebs plotted against model bleb site rank, with corresponding test distributions in lighter colours (Blue: 2% agarose, 160 blebs across 8 cells; red: 0.7% agarose, 101 blebs across 13 cells). Bins marked ***indicate where differences to the test distributions, which assumes blebs are randomly distributed, are significant on a level of 0.001 (χ² test, class widths of 1 with Bonferroni corrections for multiple comparisons). (c) Cumulative distribution functions (CDF) demonstrate that including gradients in linker strength result in much improved performance for 0.7% agarose data, with the updated CDF curve shifted up and to the left. A larger proportion of the experimental data is captured within a smaller number of model bleb sites, thus increasing the predictive power. Areas under the curve increase from 75.60% to 83.07% for the 0.7% agarose cells, whereas for 2% cells, the difference is marginal, with under curve areas of 82.37% compared to 82.48%. (d1,d2) Angular distributions of the 3 most likely bleb sites predicted by the model after inclusion of the gradient in linker stiffness, blebs are now seen to be efficiently directed towards the cell front. (91% and 95% of bleb sites in the cell front for 2% and 0.7% data respectively). (e1,e2) Cumulative frequency plots of the angular bleb distribution of experimental blebs compared to the top 3 model bleb sites both before and after inclusion of the linker stiffness gradient, for each agarose concentration.