Science: Biology: Evolution
Non-Random Mutation   (+3, -5)  [vote for, against]

We all know that evolution exists due to the mutation of the genes of successive generations, but has it ever been categorically proven that these mutations are random? Men produce new sperm all their lives, so what if this sperm is "upgraded" according to the individual's current and past experiences. So you could start out with sperm 1.0 but end up with sperm 6.4 at age 60.

So what I propose is an experiment, sequence the genome of an individual man at age 20, 40, and 60, and see what genetic changes there are to his sperm over his lifetime. Of course there will also be some random mutation, but using a sample this might be filtered out leaving the true 'advantageous' changes.
-- simonj, Jun 25 2009

Lamarckian evolution http://en.wikipedia.org/wiki/Lamarckism
[xaviergisz, Jun 25 2009]

I get the concept that mutations might not be random. I bet they are not. I bet certain stretches of DNA are more prone to mutation than others. I buy that a 60 year old will have more mutations in his sperm than the same man at 20. But then you lost me with the upgrade / experience /advantage thing. What does that have to do with mutation?
-- bungston, Jun 26 2009


The idea of a person's experiences being embedded into their own DNA is similar to Lamarckian evolution.

This idea is explored in several movies including: Resident Evil: Extinction, Alien Resurrection and The Island.
-- xaviergisz, Jun 26 2009


Bin dun.

The basic story is that mutations are random; there's no easy way to target mutations in response to, say, an environmental challenge (eg, you start drinking lots of cows' milk; what mechanism is going to "edit" the lactase gene in your sperm to make milk easier to digest for your children?). Basically, to "feed back" from environment to genome would require a phenomenally complex new system.

The slightly more complex story is that mutations are not random - some sequences are more prone to mutations than others - but the mutations are still not "directed" towards anything in particular.

The still more complex story is that **just possibly** there may be tendency for highly-transcribed genes to be more prone to mutation; so a gene that is upregulated (eg, in detoxifying a particular harmful compound) may be more likely to be mutated, duplicated etc.

Also, although I doubt anyone has sequenced sperm sequentially from the same person, the general observation is that random mutations become more frequent with age.
-- MaxwellBuchanan, Jun 26 2009


I'm more than happy with the story that evolutionary speciation is a sequence of 'relaxations' and 'focusings' - where good times lead to relaxations of what defines a viable gene pool (i.e. where random mutations are allowed to propagate - and, indeed, by promoting the survivability of the elder generation, and, in turn allowing for children to be born to elder people (with the additional chance of mutation that carries) - further relaxes the genepool as a direct result of "easy goingness" (in hard times, older parents would have long since been outperformed by their younger counterparts)) and focusing events where harsher environmental conditions trim out the fat, leaving single populations stranded at local maxima with all the less capable mutations falling catastrophically by the wayside. That twin process covers a whole lot of ground, and, as an undirected way of increasing complexity and adaptability - it's a winner.

This notion of somehow being able to encode the experiences of the body into chemical messages in spermatozoa (Which begs the separate question: What's wrong with female experience, and why should evolution disregard anything that a woman does?) seems a little magical - I mean, let's say that I do a lot of press-ups, how might my body signal that fact all the way down to my balls in order to beef-up the arms of my progeny?

There just isn't a pathway by which this information might be channelled.

And if there was, we'd be able to subvert it in ways only imagined in the most outlandish B-movies of the 60s. Which might be cool, but we'd have figured that out by now.
-- zen_tom, Jun 26 2009


//(Which begs the separate question: What's wrong with female experience, and why should evolution disregard anything that a woman does?

It's because women are born will all the eggs they'll ever have, thus leaving no chance to 'improve' them. Men on the other hand produce new sperm all their lives.
-- simonj, Jun 26 2009


Is mutation confined to dividing cells? Surely DNA damage can occur in eggs. Why else would the incidence of Down Syndrome be higher in older mothers?
-- shudderprose, Jun 26 2009


Yes, DNA damage can occur in all cells, but it's commonest at the time of replication. To be precise, any damage which is not yet repaired (as most is) by the time of replication has the potential to be "immortalized" in the daughter chromosomes (a bit like photocopying a book before you've brushed off the dead fly). So, non- replicating cells can generally fix the damage before it's too late. Also, replication itself is a hazardous process for chromosomes.

Egg cells do undergo further divisions in the ovary just before they're released, and Down's syndrome is usually the result of chromosome segregation going wrong (ie, the egg cell gets an extra copy of Chr21 at the time of the last or next-to-last cell division), and this does happen more often in older mothers. However, the number of replications is kept to a minimum precisely to reduce the risk of DNA damage, mis-segregation and suchlike.
-- MaxwellBuchanan, Jun 26 2009


Big critters like us are K-selected: good for long periods in relatively stable environments. We are in it for the long haul. Think a big company or a redwood tree. The genome is a proven blueprint that works over the long term. A given genome for a human does not exist in that many iterations in different individuals at any one time, so each working copy is precious. Given the energetic investment in the bodies carrying a human genome, it makes sense to stick close to a proven recipe. Hypermutation is too risky. Sexual reproduction and meiosis is a conservative way to stir the genomic pot and get some variation but not totally random variation: still components of a proven plan (your partner's).

For organisms that are r-selected and for which there is not much energetic investiment in a body, and which live in an environment that may be extremely variable, hypermutation has lower risk. Bacteria may grab random pieces of DNA from their surroundings and stuff them into their own genomes, on the theory that whatever let that piece of DNA loose might be something doing well in the current environment. High risk as regards whether it works, high risk for the given cell, low risk for that particular version of the genome, which probably exists in a million other contemporaneous bacterial cells. Our T cells are also in this situation. They must do battle against hypermutable microorganisms with short generation times. To match such foes, the T cells themselves have a mechanisms to cause hypermutation in their own genomes, with the end result a wider variety of T cell receptors and a greater chance of efficacy againt microbial pathogens with short generation times and high mutation rates.
-- bungston, Jun 26 2009


[Disclaimer: I have almost no idea what I'm talking about.]
What about the influence of retroviruses? Wouldn't the incorporation of retrovirus RNA into a human's DNA mean that DNA of newly produced cells would change over their lifetime? It wouldn't exactly be Lamarckian Giraffe-neck-stretching type of mutation, but it would mean the incorporation of mutations that were beneficial to the retrovirus. And one kind of mutation that is beneficial to the retrovirus could be one that benefits its human host.
-- jutta, Jun 26 2009


That would be moving from directed mutation to horizontal genetic transfer.

Horizontal gene transfer does exist, and it can be considered a form of mutation. Bacteria are mad for it, but it also happens in eukaryotes occasionally - indeed, sometimes when retroviruses go wrong. But it doesn't lead to new information (on the basis of experience) entering the system as is described in the idea. Basically it's just copying something that someone else is already doing, rather than passing on its own experience.

Within a very particular domain, some forms of 'learned' information are passed on genetically. Some microbes can shuffle their genome in particular ways to select from a choice of occasionally favourable phenotypes. They may even do this more when things look bad for them.
For single celled creatures there's no such thing as somatic cells - everything is 'germ line'. So it's much easier for them to do that sort of thing - one cell in a hundred, say, tries out a modification, and if it does well it takes over, if not, no great loss.

When you get to the human scale things are just a whole lot more difficult. Orders of magnitude more DNA, many orders of magnitude fewer individuals, much more varied environments.

I also don't like the design of the suggested experiment. If information is being passed down to then surely sequencing a man multiple times would be a waste of time - it's the gametes (sperm) which would need to change. So you might as well sequence family lines which have very long generation gaps. At the moment that's infeasibly expensive (although getting cheaper).
I'd bet good money on not finding any directed improvements, too.
-- Loris, Jun 26 2009


"Wouldn't obstruct replication, but it does give rise to an error in replication so that the newly formed DNA strand carries the mutation and you've got a virus again. But, uh, this-- all of this is academic. You were made as well as we could make you."
-- normzone, Jun 26 2009


//What about the influence of retroviruses? //

[Disclaimer - I too have almost no idea what I'm talking about, but please don't tell my employer.]

Yes, retroviruses do integrate during a host's lifetime. (Incidentally, a large part of your DNA is retroviruses that have incorporated over the aeons, and are now in various states of genomic decay). And yes, it is conceivable that the integration could benefit the host as well as the virus.

A lot depends on how the virus is transmitted: sexually transmitted viruses share the same interests as their host (ie, keeping the host fit and reproductive, or at least not killing the host quickly). I don't know of any specific cases of viruses benefitting the host, although throughout evolution this has happened many times, hence our current burden of ancient retroviral sequences.
-- MaxwellBuchanan, Jun 26 2009


[simonj] you repeated my point - yes, your experiment is looking for a male-centric directed mutation vector that supports a Lamarkist point of view - however, if this is the case, then you have to accept that this process is not available to females. My (unspoken) point was that it seems to me that women (and females of all sorts of species) have evolved all sorts of adaptive features. If women can evolve too, then where does that leave your theory?

Aren't retroviruses "just" another vector for mutation though? And again, if their Lamarkian action could only be passed down through male experience, mightn't that skew the resultant population? [I don't know what I'm talking about either for what it's worth]

I did hear somewhere that a person's genetic 'library' of antibodies might be influenced by the bacterial (and viral?) exposures that their ancestors underwent, suggesting a directed adaptation in that respect - but I don't know what the process was.
-- zen_tom, Jun 29 2009


//you could have a mechanism// I think you've just described exactly how evolution works, in signal processing terms - //dumb generation and smart selection// is precisely it - I don't see how it could sensibly work any other way.
-- zen_tom, Jun 29 2009


[Disclaimer - Not only do I have no idea what I'm talking about but I also have no idea what any of you lot are talking about either!]

//Of course there will also be some random mutation, but using a sample this might be filtered out leaving the true 'advantageous' changes//

How do you determine what is an advantageous change? Assuming for a moment that changes in a single individual happen often enough for this to be a worthwhile experiment, I can see that some of the more obviously disadvantageous changes might be easy to determine, but how would you be able to decide whether a change provided an evolutionary advantage?
-- DrBob, Jun 29 2009


//Aren't retroviruses "just" another vector for mutation though? And again, if their Lamarkian action could only be passed down through male experience// Yes, retroviruses are a source of mutation, but there is nothing Lamarkian about it in the normal sense. To all intents and purposes, they generate random mutations (from the host's perspective), and natural selection favours the tiny minority of ones that happen to be favourable.

//a person's genetic 'library' of antibodies might be influenced by the bacterial (and viral?) exposures that their ancestors underwent// Antibody repertoires are very different between individuals, and this variation is subject to selection like all variation. So, if the population suffered a pandemic disease, then antibody genes that can protect against that disease will be more abundant in the survivors; again, not Lamarkian, just selection.

(As an aside, antibody genes get shuffled and mutated within the individual to further adapt to environmental challenges. But again, it seems to be a selective rather than Lamarkian process - there's a mechanism for mutating the genes in the right cells, but there isn't a mechanism that says "Ah - here's a new antigen; if I change this "T" to a "C" it'll improve the antibody's affinity for the antigen.")
-- MaxwellBuchanan, Jun 29 2009


//how would you be able to decide whether a change provided an evolutionary advantage?//

Easy. A prospective randomized study over the next 500,000 years.
-- shudderprose, Jun 29 2009


The proposed experiment is unclear in how to differentiate 'random' from 'advantageous'...
-- loonquawl, Jun 29 2009


[MaxwellB] I remember reading somewhere towards the end of Dawkin's "The Extended Phenotype" some theorisation about how some processes in the body might propagate themselves (and by extension, their phenotypes) through host generations either directly through DNA recombinations, or indirectly through mother/foetal mechanisms.

In short, the idea was that there is a lower level of 'evolutionary' activity that occurs within a body's lifetime (between, for example, competing antibodies) which might be conflated, during gestation, with a body's genetic development and phenotypic 'identity'.

Dawkins wasn't suggesting that it happened, but that there might be a mechanism - I forget the details, but there was talk of the environment created within antibody, bacteria and viral interactions that might be passed down, and indeed have effect upon the expression of the genetic line.
-- zen_tom, Jun 29 2009


Hmm - don't remember that bit. Was he talking about epigenetics (ie, regulation of gene expression)? There are certainly stable changes in gene expression, and some evidence that these changes can be passed on to offspring, a phenomenon known as imprinting. But I'm not sure what Dicky was on about (and I'm not an expert on genetic imprinting either).
-- MaxwellBuchanan, Jun 29 2009


<looks in vain for an emoticon that represents the passing of a flattened hand backwards over the top of the head, accompanied by a 'whoosh' noise> Exactly. <\livfaetrtpoafhbottoth, aba'w'n>
-- MaxwellBuchanan, Jun 29 2009


Also:

livfaetrtpoafhbottoth, aba'w'n [marked-for-tagline]
-- MaxwellBuchanan, Jun 29 2009


(marked-for-tagline)

"promulgate language permutation faster and more and sexier"
-- normzone, Jun 29 2009


that's some powerful peyote there [Ian]. Most of us prefer to stick to the "this is how we're pretty sure it works" semantic. mode. Since this post doesn't posit any invention, mode mechanism or discovery therein I will take the apropos position of [marked-for-deletion] WIBNIfty
-- WcW, Jun 30 2009


I suspect that the female line is just as able to engage in nonrandom mutation as the male line. Possibly more so in fact. Just differently.

While the female eggs are all meiosed at the same time, and from then on, the body itself doesn't do much to mutate them, transcription errors made by the body are not the only sources for mutations.

Cosmic rays, as you may be aware, pass through the body all the time, and should one hit a sex cell in a woman, there's every chance it will alter a gene.

While we believe that the order in which ova are shed from the ovaries is random, it may well be that the body is selecting eggs which stand a greater, or lesser chance to have been mutated (either based on location, or on chemical evidence of alteration) depending on the apparent situation outside. Similarly, the period between uterine wall sheddings could be altered based on how favorable the body feels towards a specific egg.

As water is one of the better shields against cosmic radiation, it may also be that a woman's ability to retain water, and not to retain water relates to the body's "desire" to achieve mutations in the ova.

You'll note that here the mutation is random, but the attempt to pass mutated cells on is not.

Am I aware of mechanisms the body might use to regulate these activities? Uhh, no. However, we could do some complicated studies to find correlations between stress levels, water retention, location of ova release, and time between periods.
-- ye_river_xiv, Jun 30 2009



random, halfbakery