Developer diary 2

1 Oct 2009

I've put this email separately as it's a bit longer. Later I'll add the graphs as part of the analysis of the simulation.

15th February 2005

I've attached the promised graphs. I haven't labelled them (GCSE science sin) so this is what they are. The first shows the time it takes for the eight fastest grower of the population to go from 5 to 50 cells. I didn't put all the mutants of the generations there because the earlier generations have fewer cells than the later ones (16 compared to 64) and some of the mutants don't grow at all. As you can see the increase is pretty much linear for the first ten generations at least (it works out at roughly 500 time units increase per generation). However, they seem to reaching the maximum growth rate by Gen 13 and the population is getting more homogenous. They just about managed to break the 10 000 mark. Hooray!

Growth time per generation for heterocyst simulation

(It would probably have made more sense to swap the axes and show how growth time decreases as the generations pass).

The second graph shows the (nicely exponential) growth rates of the fastest growing mutants of every other generation (to make the graph clearer). I'm not sure what the best way to compare growth rates is. Adding an exponential trend line (final graph, which I wasn't going to add), you can see the first mutant (the ancestor) grew with an e to the 0.1619(time/1000) relationship. The best grower grew with an e to the 0.2396(time/1000) relationship. So what does that mean? Would it be right to say that the fastest grower is 48% faster (0.2369/0.1619)?

Number of cells over time for various generations of my heterocyst simulation

(By the way these are screen shots of the graphs rendering in Google docs, which is why they look a bit weird.)

Fitting exponential curves to the fastest growing heterocyst filament from the first and thirteenth generations.

I've also started to analyse the mutations in the fastest grower (so this is where all my time's been going). I added them one by one to the ancestor to see the effect on growth. Most have some small effect, a couple actually make it grow slower, and one kills the whole filament, which is nice, because it shows how the mutations are dependent on each other. The one that kills the filament actually reduces the degradation of a certain enzyme thus massively increasing the concentration it reaches. The ancestor cell can't cope with this increase because it doesn't have a good enough regulation system to deal with it. This mutation occurred a few times, but died out. Then it occurred in the 7th generation (in 17th mutant of this generation to be precise), which could cope, and this individual was the fastest grower of the generation (it's about 1000 time units faster, as you can see on the first graph), and so spread the gene throughout the subsequent generations until everyone had it (unless they mutated it again, in which case they were much slower). I've also noticed that the enzyme concentrations in heterocysts are not stable as in the first organisms but oscillate quite a lot. Quite why this is or how it helps I don't know. So much to find out. So much more to try.