T cells stimulated in a well for 24 hours.  

In silico, each T cell is an agent that obeys simple, but stochastic,
rules that depend on its attributes and the cells in its
neighbourhood.  Each cell has x and y coordinates and a radius.  Each
live cell, independently, has a constant probability per unit time
(rate) of suffering death.  Each also has probabilities per unit time
of becoming activated and of entering the cell cycle, that depend on
the number of nearby IL-2 producing cells and CD25+ cells. Cells do
not move but they increase their radius on activation. Thus, spatial
inhomogeneities in the computational model are the result of cells
interacting with each other.

In vitro, we make use of the simple, non-disruptive observational
method of taking phase-contrast photographs of cells.  The images show
inhomogeneities, with dark patches forming and growing, consistent
with activated cells being able to activate and adhere to neighbouring
cells.

1. Supplementary .pdf (to see the animations, open with acroread).

Above: In silico realisation of 50000 interacting T
cells. Heterogeneity arises from local influences such as the
proximity of IL-2-producing cells (red).

Below: Photos of a single well containing CD4+ T cells from a BALB/c
mouse, under stimulus with PMA and ionomycin.

2. Animated .gif. Photos of cell culture.

3. Animated .gif. In silico realisation.

4. Python code.

Cells are allocated positions and attributes before timecourse begins.
Code will run as anaconda distribution python2 or python3.
The initial number of cells is set to 20000. Smaller (larger)
values will give shorter (longer) running times.