Currently, it's possible to weakly assign colors to functionality by changing the simulation settings for specific colors. However, this is only a suggestion and doesn't enforce anything and introduces a few issues.

For example:

• The cell can be any color still, it just won't function if it is assigned to a color of which the cell type doesn't have any value
• The organisms might have lots of cells that end up doing nothing because the correct color is statistically unlikely to be mutated to
• This system doesn't factor in cell types that don't have settings you can modify per color, such as neurons

So my suggestion is to introduce a different system; color specialization. Here's how it would work:

• Each cell type has an option in the settings to select which colors it can exist as. For example, you might only want attacker cells to be red. You would then select the color red as the only color an attacker cell can be.
• Cells that are of a type that is not assigned to the color they are simply cannot exist
• The color mutation will only mutate between colors that a cell type is allowed to use, for example if you want two colors to have slightly different functionality for the same cell type
• The cell type mutation will simply change the color of the cell, given that the cell is able to mutate to that color via the cell color transition matrix.

This creates an enforcement of cell functionality with colors, which would allow more intricate visual patterns being developed as an organism evolves. It also ensures that cell types that are of a color which have no functionality from the settings (for example a green attacker cell with attack strength set to 0) do not happen and therefore the likelihood of functional cell networks evolving is higher.