I want to create evolutionary simulation on the chemical level. Life is essentially self-replicating chemicals, which is what I want to simulate. How would mechanics of such system work? In other words, how does one get self-replication through chemical-reactions?

Firstly, you might want to see how DNA and RNA molecules work. That would probably give you some ideas.

I looked up DNA and RNA on wikipedia and in a textbook.

The gist of it is that DNA and RNA are instructions/programs.

But they fail to tell me how these instructions are read/executed - which is the most interesting part.

Well, the instructions are coded in the form of nucleotide sequences. Each sequence of nucleotides (if i spelled that correctly) on a DNA is capable of producing a corresponding RNA, which can then be used to create certain proteins. It actually works kind of like a jigsaw puzzle.

We know there are 4 distinct elements types on a strand of DNA, named A, T, C, and G. They have certain pairings too. I think it was A with T and C with G or something like that. So, because they can only fit together in one way, that makes it possible to use 1 strand to construct a complementary strand automatically. We know that a DNA has 2 side, and that is its stable state. When RNA production occurs, a segment of the strand breaks apart and complements of that segment is created, which then is the RNA instruction. (Hope I'm remembering this correctly.) That strand of RNA then goes off and is then used to create a protein sequence. Oh yeah, one more thing, the pieces aren't really "constructed" in our normal sense, they just kind of come together because there's only one way things will fit.

So, let's assume that the pairing sequence I mentioned above is correct and we have a segment of DNA with the sequence AACGTGCCA, which of course is only one side, but we know what the other side is because there is only one way pieces fit together. So, based on this strand, an RNA can produced, which will be its complement, TTGCACGGT. The RNA then goes off to be used to produce a specific protein. The protein production happens in the same way. The RNA then becomes like a mold in which amino acids bond to and find their place in the sequence. Once the protein is complete, the RNA breaks away.

So, in a sense, chemical reproduction is not "reproduction" in the normal sense. The replication process requires some sort of mold or a catalyst, which brings elements together. The catalyst can either act as a mold or act as a source of energy to "persuade" the elements to bond.

Hopefully that was slightly helpful.....

Interesting... what do the proteins do?

KulSeran said:

then each block of 3 encodes one of 23 diferent protens.


Actually, each block of 3 nucleotides (a codon) codes for a specific amino acid, and there are only 20 found in proteins. The other 3, citrulline, ornithine, and ?, are not found in proteins.

One of the best examples of this type of system that I've seen is JohnnyVon. You can access various papers and versions 1.0 and 2.0 of JohnnyVon at http://johnnyvon.sourceforge.net/ Version 1.0 seems to me more DNA-like, while 2.0 looks more like aggregation rather than self-replication.

I had a program in mind, CORE_WARS: http://en.wikipedia.org/wiki/Core_Wars

My problem with CORE_WARS is that it's 1-dimensional, non-evolving, and relies on an interpreter to read/execute the commands (instead of self-emergent behaviour through chemical-reactions).

But I DO like the idea of virus-bacteria that gradually evolve to greater levels of complexity.

Are there any programs like that?

Yeah, evolution isn't real. But then who said games had to be realistic anyway?

Yeah, but I'm not making a game... I wanted to make a simulation.

I ran a few simple simulations with lots of interesting results, so I wanted to make something more advanced.