?

Log in

No account? Create an account

Dr Who: The Lazarus Experiment - John C. Kirk — LiveJournal

May. 8th, 2007

01:00 am - Dr Who: The Lazarus Experiment

Previous Entry Share Next Entry

Comments:

[User Picture]
From:susannahf
Date:May 8th, 2007 11:00 am (UTC)
(Link)
If you believe in the premise of "The Selfish Gene" - which is reasonable but has some flaws, there is also an evolutionary reason for aging.

Assumptions:
1) Organisms exist to perpetuate their genetic information
2) Genes that confer advantage to offspring will dominate

From this you get the conclusion that the most "successful" genes will cause their carrier organisms to be highly fertile and resistant to disease. However, there are limited resources, so you can't be too fertile (otherwise all your offspring will die of starvation), and you shouldn't have infertile organisms (old people) using up resources that could be better spent on fertile or pre-fertile organisms.

So aging can be argued to be a desirable trait from an evolutionary point of view. Once you are no longer capable of passing on your genetic material, you die.

Of course, this is a gross simplification leaving out all sorts of aspects like post-fertile adults providing childcare and education while the fertile adults do the hunting and so on, but the gist still holds.
(Reply) (Parent) (Thread)
[User Picture]
From:karne_k
Date:May 8th, 2007 01:19 pm (UTC)
(Link)
I'm trying not to side-track the discussion too much, so I'll keep this short :)

>From this you get the conclusion that the most "successful" genes will cause
>their carrier organisms to be highly fertile and resistant to disease.

Not quite. The Selfish Gene says:

1 - genes exists to replicate (and replicate to exists)
2 - genes replicate via the mechanism of animals (plants, bacteria, slime moulds etc.)
3 - the most common genes are those that best help their animal pass on more copies of themselves

The latter means that the 'best' genes are those that best help their animal (i) survive within their ecological niche (or take over another one) and (ii) produce the most offspring *which in turn must survive to breed*.

This implies that diseases that occur after reproduction are irrelevant to evolution (although not the case with primate or other animals with significant levels of 'grandmother' infant care) and that high fertility is only required in situations of high infant mortality. Small numbers of well cared for offspring, whom in turn have an excellent chance of passing on their genes to their own offspring, are actually more valuable from the selfish genes' 'point of view' than large numbers of poorly/not cared for ones, many of whom die. Logically, the former strategy is aided by longer life spans and higher levels of intelligence. Cooperation between related individuals, group living, communication and the development of language also all help. Think pride of lions, troop of baboons or humans societies.

> However, there are limited resources, so you can't be too fertile

Efficient use of limited external resources isn't really an evolutionary issue per se, its effect is minimal unless an animals' breeding is so extreme that they use up all the resources in their niche before their offspring themselves can breed. The selfish gene isn't able to see the long term picture (please ignore my anthropomorphic terminology!), otherwise we'd not see the common cycle of 'boom-bust' in animal populations.

Dragging myself back to the point. I'm sure aging does have an evolutionary aspect, but I suspect that it's more a case of (i) how much internal energy/resources should be put into this individual breeding now vs living longer? and not (ii) when should this individual die off to allow their offspring access to its external resources?

As always, there are some special cases – spider mothers using their own bodies to feed their young, being one I can think of. In general though, the most powerful evolutionary forces are the simple ones – the ones that work directly on individuals and directly effect their breeding success.
(Reply) (Parent) (Thread)
[User Picture]
From:johnckirk
Date:May 8th, 2007 02:58 pm (UTC)
(Link)
I remember a line from a Spider-Man comic, where the High Evolutionary said that "cloning stagnates the evolutionary process". So, from the point of view of improving the species through natural selection, I'd agree that you don't want immortality for everyone. (I think it was one of the Dawkins books which said that evolution relies on non-random death.)

Thinking about genetic algorithms, the idea is that you start out with a pool of possible solutions, and then use them to generate new ones. However, unlike biological organisms, the old solutions don't automatically die out - they get compared to the offspring, and then the best subset from both generations combined gets to live on. So, there's no real concept of age: you may wind up with an original solution still intact after 70 generations, and it's just assessed on its own merits.
(Reply) (Parent) (Thread)
[User Picture]
From:karne_k
Date:May 8th, 2007 03:57 pm (UTC)
(Link)
[posted again, sorry John please nuke the anon post - I hate LJ sometimes! :) ]

Evolution is neither good for the species (which is an arbitrary descriptive label used by humans because we can understand it better than 'semi-closed subset of the available gene space') nor any individual in it. It fact it's not good for anyone; it's a natural process that occurs whenever you take information and process it in a 'genetic' fashion. You can see evolution occurring in e.g. software viruses and antivirus systems. The concept of the rightness of the 'March of Evolution' is a religious idea as much as any other. Humans have used evolution as a tool for millennia, whether it be for animal husbandry or microchip design - seeing it as anything else is silly :)

>However, unlike biological organisms, the old solutions don't >automatically die out

Don't relate a genetic algorithm's solution to an individual. A better relationship is between each separate test that you run on that solution (to check its efficacy) and an individual. You'll probably run hundreds if not thousands of tests (in order to get statistical strictness) and the same applies in nature. Old gene solution certainly don't die out - not at the time scale you're thinking of and mixing between the old and the new is very common. Indeed, the concept of generations of solutions in genetic algorithms is artificial (and imposed simply to make things easier for us humans); evolution in nature is a lot less quantised than that.

As in your example, if you take a culled population of say 1000 rabbits of a particular genotype and you put them through 70 generations (oo.. say a 50 years) in environment that they've evolved to match and if that environment doesn't change, then I'd fully expect the rabbits at the end to be very similar to the ones you started with. Unless they’ve evolved to avoid your culling technique, of course.

You'll also have one hell of a rabbit skin carpet...

(Reply) (Parent) (Thread)