Skip to main content
. 2018 Nov 20;6(1):1800927. doi: 10.1002/advs.201800927

Figure 1.

Figure 1

LRRK2‐G2019S dependent phenotype was enhanced in 3D conditions. A) Schematic representation of the experimental pipeline used in the study. Image acquisition, segmentation, feature extraction, and data analysis were all automated. B) Amount of TH+ pixels in 2D and 3D cultures after 6w of neuronal differentiation. Values represent means with whiskers from min to max (number of wells in 2D: LRRK2‐WT 48, LRRK2‐G2019S 39; number of bioreactors in 3D: LRRK2‐WT 48, LRRK2‐G2019S 34). C) Representative maximum intensity projection of confocal images of LRRK2‐WT and LRRK2‐G2019S neurons after 6w in 2D showing TH mask after segmentation (red line) superimposed on TH raw channel. D) Representative confocal images of TH+ cells and consequent TH skeleton mask and branching rendering after 2w. E) Radar plot showing several features extracted from TH segmentation after 2w (number of bioreactors: LRRK2‐WT 31, LRRK2‐G2019S 32). F) Representative confocal images of TH+ cells and consequent TH skeleton mask and branching rendering after 6w. G) Radar plot showing several features extracted from TH segmentation after 6w (number of bioreactors: LRRK2‐WT 48, LRRK2‐G2019S 34). Scale bars 100 µm. In all cases, p‐values are calculated using Mann Whitney test and they are adjusted (red) or not (black) according to Benjamini–Hochberg (number of total features assed 18), *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. For the 3D analysis, 420 fields at 20× magnification were acquired for each bioreactor (21 fields for 20 planes). For the 2D analysis, 45 fields at 20× magnification were acquired for each well (15 fields for 3 planes). Each data point corresponds to a bioreactor.