Developer diary 1

15 Feb 2010

Because I created my heterocyst simulation a long time before I got a blog, I can't write about my progress in creating the simulation. However, I have collected some fragments of some emails I sent a friend while I was working on the program. Some of the analysis I write about, and include in later updates. I also found a brief history of the program at the top of the code. As you can see it took me quite a long time to get anywhere.

# Jul 04  Got working with simpler reactions
# Aug 04  Added enzyme kinetics and diffusion
# Oct 04  Added protein degradation and transcription factors
# Jan 05  Tidied enzyme and metabolite variables; added replication
# Jan 05  Changed regulation system and added (im-/ex- portable) genetics
# Feb 05  Integrated crossing program for evolution
# Feb 05  Changed regulation to consensus system and separate inhibitors

July 2004

Wikipedia has no results for heterocysts. I'm tempted to write one later.

I did write one later; it's here. It remains one of the few articles that I've created from scratch.

20th January 2005

Those damn heterocysts kept me up until passed 4 the other morning. I can't remember why, but it probably had something to do with avoiding work, that I decided to look at my heterocyst program again and mess about with it. I managed to sort a few things out and of course thought of hundreds more things I should do...
... just as I hope to evolve some heterocysts some day.

30th January 2005

I finally got my heterocysts to replicate and spontaneous form along the growing filament at regular intervals. It's mesmerising to watch them appear, which is why I've spent most of the day staring at a screen of numbers and coloured boxes with excitement. I've also managed to generalise lots of variable for the regulation system and so I've started a little evolution. Of the fifteen mutants of my original heterocyst, three performed ever so slightly more efficiently (grew faster), and loads performed terribly. I think it will take hours to get any really nice results, especially as you need quite a large population for any decent evolution. And loads of generations.
It would be nice to start with no regulation system and try and evolve it, but that would take a long time.

7th February 2005

My heterocyst evolution program is becoming more and more autonomous, so I can leave it evolving all day while I am at lectures. And they're doing well. It's amazing how fast they seem to improve. I wonder if they'll reach a limit some time soon.

8th February 2005

I may email you stuff about heterocyst evolution that I have found. Last night, there was a great, super fast mutation, but after the program had been running for hours and I stopped the program to copy the data to Excel, I accidentally ran the program again, re-writing all the data. Man, I was annoyed. Fortunately, I have since got a mutation that I think is as fast, and if not another generation should do it. I suppose in all of evolution fit individuals have been unlucky (struck by lightning or whatever), so this just mimics that.