D. Förderung des wissenschaftlichen Nachwuchses

Epithelial cells: Dynamic heterogeneity is linked to wound healingperformance
Last year we described results of preliminary experiments that examined the
wound healing performance of cellular cultures at different densities. The cel-
lular density strongly affects the strength of dynamic heterogeneity (SDH) - at
low density, there is low SDH because most cells are highly mobile, while at high
density there is high SDH because some cells are mobile while others are jammed
in place due to intercellular interactions. In the past year we have completed this
study by incorporating cellular shapes, diffusivity, and intercellular coordination
measures into the analysis (Vishwakarma et al. 2020a). These results show that
SDH explains wound healing performance due to differences in cellular coordi-
nation: Cell cultures at low SDH have lower intercellular coordination and are
slower to heal a wound, while cultures at higher SDH (which includes the typical
homeostatic density of healthy cells) have higher intercellular coordination and
are therefore faster to heal a wound. Furthermore, reflection interference con-
trast microscopy (RICM) was used to reveal cell-substrate adhesion points and
show how cellular density affects the orientation bias and strength of adhesion
of leader and follower cells. We summarized these and other related findings in a
review article that discusses what is known about intercellular signaling, mecha-
nical forces, and physical heterogeneity in relation to leader-follow dynamics du-
ring wound healing (Vishwakarma et al. 2020b). The combination of these factors
suggests that the motion and coordination mechanisms in epithelial tissues have
evolved to attain a mechanical resilience that enables a regulation of homeostasis,
yet also permits a quick switch to cohesive collective motion for efficient wound
healing.
Honey bees: Unsupervised characterization of behavioral variability
We use barcode tracking of age-matched cohorts of honey bees to characterize
both behavioral patterns of individual bees on a given day as well as changes over
a bee’s lifetime. We used an unsupervised data-driven approach to describe beha-
vioral differences. This approach involves first defming behavioral metrics that
describe both where and how a bee moves within the nest. In contrast to other
works that assign categorical descriptions such as, brood-care bees, nest workers,
and foragers, we use an unsupervised approach (dimensionality reduction and hi-
erarchical clustering) to quantify and describe the full ränge of observed behavior
Variation. The dominant axes of Variation are whether a bee is engaged in foraging
Figure 1. Unsupervised description of honey bee behavioral variability on agiven day. The diagram ►
shows a t-SNE embedding oj the data, obtained used behavioral metrics calculatedfrom motion trajectories of
4000+ individual bees tracked over a 50-day experimental period. The colored clusters highlight differences in
behavior, shoiving the distribution oj behavioral metrics and average nest location histograms. A representative
diagram of nest contents is shoivn at the lower right for reference.

300