pouët.net

genetic modeling

category: offtopic [glöplog]
 
I'm currently listening to this book. As someone who used to write "3d modeler" into his C.V. the concepts shown there blow away. What Sean B. Carroll describes sounds like a manual to a procedural 3d-modeling tool: Complete with selections, boolean operations, extrude, and replicate.

It should be straight forward to convert his gen-toolkit-descriptions into a basic set of operators that control a forward particle simulation (of cells). Did somebody else here thought about that idea?

Maybe I'm a little bit too enthusiastic, but I'm pretty convinced that I'll probably see that "genetic modeling tool" being implemented in my life-time. Damn, I once again wish I would have picked another career.
added on the 2012-11-22 23:58:25 by pixtur pixtur
http://picbreeder.org/
http://endlessforms.com/

Using an evolutionary approach tends to work better if the change introduced by the random variation have a high likelyhood to have a limited impact, ie. not playing directly with an operator tree. Those 2 examples use a different approach, they work with a scalar field and a function representation crafted specially for an evolutionary approach.
endlessforms.com is pretty cool. But the random-selection approach feels to in-precise.

I have to dig into the pic-bender, once I have a little time.

I'm still looking for an easily digestible list of the toolkit- and switch-genes Carroll is describing in his book, together with a breakdown on how and in which resolution they're stored within the DNA.
added on the 2012-11-23 06:09:30 by pixtur pixtur
AFAIK Christiane Nüsslein-Volhard was awarded the Nobel Prize for the discovery of some genes that trigger developmental processes.

I guess, though, that a list of such genes would be very long (and incomplete).
added on the 2012-11-23 11:56:32 by Adok Adok

login