Figure 2 : Deformation frequency and PC1 of Saos-2 cells on 10 surfaces tested.

From: A high throughput approach for analysis of cell nuclear deformability at single cell level

Figure 2

Two descriptors, circle variance and rectangularity were used to characterize the extent of deformation on the surfaces. (a) Based on the test data, five regions were identified in the rectangularity-circle variance space that quantified the degree of nuclear deformation. Among these, regions 1 and 2 (cut off: circle variance, rectangularity ≤0.2) represented minimal deformation. After constructing circle variance versus rectangularity plot, the region with the most circular nuclei was labelled as Undeformed and rest as Deformed region. For each surface, fraction of the nuclei falling in Undeformed and Deformed regions were calculated. Control surface had the highest (77.5%) Undeformed nuclei fraction while P4G4 had the lowest fraction (8.60%). (b) Principal component analysis was conducted for circle variance and rectangularity descriptor measurements and Principal Component 1 (PC1) results were further analysed statistically (*p < 0.05, ns: non-significant). Distributions of cell nuclei PC1 were not normal according to Kolmogorov-Smirnov test so Welch’s ANOVA with Games-Howell Test was used (p < 0.05) to show significance. All nine patterned surface shape distributions were statistically significant compared to the Control surface. Also distributions of surfaces with different gap sizes were found to be significant. In the 4 μm gap group (P4G4, P8G4, and P16G4) only the distribution of P4G4 was significantly different than the other two, and among all the surfaces, P4G4 had the highest significant difference with the Control (μx control = −0.150648, μx P4G4 = 0.084128).