C6 is a toy center-based model of the initial clustering that occurs as a cells aggregate after replating.
- Represents cells as non-overlapping circles
- Cells exclude each other
- Cells sense each other with tunable strength and distances
- Cells grow at tunable rates
- Cells divide at tunable sizes
- Cells' growth is contact inhibited
- Runs are launched from stored initial conditions
- Runs are selectable as deterministic or stochastic
- Runs are logged to tidy data file
- Run state is visualized
- You can save run states out to pretty MP4s
- Summary statistics are available for probing run logs
- Easy run profiling
import c6
import numpy as np
import matplotlib.pyplot as plt
space = c6.Space()
cell_locs = 20*np.random.random((10,2))
cells = [c6.Cell(space, loc) for loc in cell_locs]
fig, ax = plt.subplots(1,1, figsize=(10,10))
ax.set(xlim=(-10, 20), ylim=(-10, 20))
animation = c6.plot.animate(fig, ax, space, 500)
animation.save('example.mp4', fps=20)
Clone and install or pip install git+https://github.com/AllenCellModeling/c6.git
Both initial conditions and run logs are stored as JSON in this form:
{
"seed": 123,
"universal": {"sensing": 12.3,
"influence_max": 12,
"influence_decay": 0.3,
"adhesion": 0.5
},
"cells": [{"time": 0,
"id": [0, 1, ..., 100],
"loc": [[1, 2], [3, 4], ..., [0.5, 2]],
"radius": [0.3, 0.5, ..., 2.0]
}]
}
Any property that can be passed to c6.Cell
as a parameter can be serialized either as a key/value pair in 'universal'
, in which case it applies to all cells at all time points, or to "cells"
, in which case it is parsed on a per cell per time point basis. Starting conditions are this file, but with only the first entry in "cells"
specified. The first entry in "cells"
is always applied before later entries and so per-cell values that do not change (e.g. ids) can be set there.
For full package documentation please visit AllenCellModeling.github.io/c6.
Available under the Allen Institute Software License