Aggregation dynamics in heterogeneous cell populations – ADHeC

The organization of multicellular assemblages derived from an aggregation process is impacted by heterogeneity among the constituent cells. Affected properties include stability, size and composition of groups. Differential assortment can result in disruptive competition within aggregates and lead to the long-term failure of collective functions. Models that describe mechanistically the emergence of collective behaviour from cell-level properties, however, typically consider homogeneous populations. This project will address aggregation in binary mixes of cells with controlled differences in microscopic properties. This setting offers a simple implementation of heterogeneity, yet one that is relevant to many contexts, e.g. in the presence of a mutant sub-population.
Complementary theoretical approaches – individual-based, self-propelled particles and macroscopic PDEs under maximum density constrains – will be used to describe the qualitative dynamics of the aggregation process. The fraction of cells of one type in the population will be used, along with a metric of the individual differences between types, as a control parameter for the exploration of the phase space and the classification of aggregation patterns.
Simulations will be benchmarked against experiments realized with an organism that naturally forms heterogeneous cellular aggregates: the 'social' amoeba Dictyostelium discoideum. We will study the aggregation process and the successive developmental fate of binary chimerae, where we can control the initial population composition, the nature (genetic or physiological), and in certain cases the magnitude of cell-level differences. We will explore the role of microscopic heterogeneities, notably in cellular movement and adhesion, in determining population-level patterns. Direct measures of cell behaviour and knowledge of its genetic underpinnings will help us connecting the observed patterns to the microscopic differences between sub-populations.
Machine Learning will be applied to establish a map between microscopic and macroscopic representations first, and, after training with synthetic data, between models and observations.
Models for aggregation will be moreover introduced as ecological modules in evolutionary frameworks to predict what types of heterogeneity, through their effects on population partitioning and segregation, sustain or disrupt collective function on a long time scale.