AI
Friday, November 28, 2003 7:45:55 PM
You raise good points, DougS. One big difference between computer programs and genomes is that computer programs are highly linear, "one sequence has one effect" sorts of entities, whereas in my basic understanding, genomes reuse code, proteins, enzymes, and other mechanisms redundantly for all sorts of related and unrelated purposes. So in that sense, things are vastly more complicated in a genome than in a computer program. A critically-located error in a genome can cause a cascading array of symptoms and diseases; or, because of all the redundancy, have very little effect unless several other critically-located errors crop up too. I've read about both types of genetically-based disease, and I know of similar things happening in computer programs, although in programs things tend to go critically bad much, much quicker. 
I still don't think the analogy in and of itself is off, but the scopes are much different, as you point out (and as I suspect mingwan0 is likewise pointing out). Even with the larger scope of the problem, though, more progress will be made accounting for it than by pursuing an incomplete or incorrect theory. I still think it's a good thing that scientists are acknowledging the relevance of noncoding DNA--and if DNAPrint's maps cover them already, then we're ahead of the game right now!
--Nik
I still don't think the analogy in and of itself is off, but the scopes are much different, as you point out (and as I suspect mingwan0 is likewise pointing out). Even with the larger scope of the problem, though, more progress will be made accounting for it than by pursuing an incomplete or incorrect theory. I still think it's a good thing that scientists are acknowledging the relevance of noncoding DNA--and if DNAPrint's maps cover them already, then we're ahead of the game right now!
--Nik
