Evolution: time for some change?

this might be pointless because development is highly modular. functions can be adapted and switch and to accomplish the same structure different modules can be used in quite related species. I was looking myself at the stem cell compartment of the tooth in mice and voles and noticed for instance that the WNT ligands (one of the major signaling families) are not expressed similarly. This project has never reached an advanced status but possibly different WNTs take over the function of another WNT in the same tooth in vole and mouse. Or because I actually did not find any subsitute for certain patterns yet, another signaling pathway may have taken over the role completely. And that to create a tooth which is the same in shape and function in both instances. This is just a personal example, but I am sure it is not the only one.

What is therefore then the point in looking at elemental building blocks of genes? They don't necessarily exist in this fashion. A similar functional system can arise by different developmental means, using different genetic 'blocks'.

 

Log in or Sign up to hide all adverts.

Thanks for the refs. They look interesting ... I should get a chance to read the papers sometime next week. A bit busy this week (presenting my thesis proposal tomorrow to my advisory committee). My name's Mark as well ... but I haven't published anything I'm too proud of, let alone relevant.

Maybe I should have emphasized "fit together into functional systems". One certainly has to evaluate things on the network or system level, but how the components fit together is critical for this purpose.

Presumably there are a large number of ways a system can be organized while maintaining the same function (just as you say). All networks with the same function can be called instances of a single "meta-network". In evolution you move both between instances of a meta-network and more interestingly, you occasionaly move between functionally different meta-networks.

If one can count the number of instances of two meta-networks (or estimate their size somehow) and the number of paths between specific instances of each of the two meta-networks (under some survivability constraint), then one will begin to understand the evolution of such meta-networks. You definitely need to sample the networks at a reasonable scale because the numbers are combinatoric and there is a large flexibility in the number of ways instances can occur.

This of course is very experimentally laborious. Particularly depending on the level of resolution with which one understands the network (quantitatively vs qualitatively). Using modeling and sampling techniques would certainly be helpful, but it's unclear one could make good models at this point. Therefore one needs both research into understanding the molecular details of example systems, like Eric Davidson, and sampling of systems in related organisms (of which you seem more aware).

 

Log in or Sign up to hide all adverts.

Doesn't evolution happen at the level of populations? Once you're an organism, your evolution's done.

Except in Star Trek, where people evolve spontaneously.

"Hey Sam, wanna go out?"
"Nah, I thought I'd stay home and evolve."

 

Log in or Sign up to hide all adverts.

Every organism is a data point in the statistical sampling that leads to improvement at the population level.

The reproductive success of each individual is a measurement that leads to information at the population level in terms of what works well.

You require a reproductive chain of individuals to evolve - it's not like individuals bud off of a population and then the two go there separate ways. Each individuals contribution back to the population is very important. They aren't two separate entities.

 

That is so, but individuals cannot EVOLVE, they can only be SELECTED UPON. Generally when you're a multicellular organism, genetic adaptation only happens once - when you're conceived.

Everything after that is about environment, luck, and the success of your design. If you chance to reproduce, then your offspring also get to adapt - once. Evolution is really a name that humans apply post hoc, when some pressure is put on a population and some of them don't die.

I agree that creating metanetworks describing functional components of organisms would be experimentally laborious - that's probably an understatement. The workings of the nuclear DNA itself, with inhibitors and activation sites and so on, is fiendishly complicated, and I understand that the protein interactions which result from them are vastly more difficult to understand.

The biggest problem, most likely, is that our understanding of organic genomes (our own, for instance), is pretty bad. People still refer to non-coding regions as if we were sure that they didn't do anything... if you go on a gene-by-gene basis we share most of our genetic code with chimps, and also with blowfish and probably some species of mold. Trying to break down an organism into all of its functional components (including those that aren't expressed, but can be with minor mutations, like the human sixth finger / gills / other "junk DNA" like that) would be an Herculean task even for simple organisms.

 

Okay, I see you're working off a direct quote - I meant selection occurs at the level of the organism. My point was to say evolution selects at the level organism not the gene.

Although I wasn't claiming it - it is easy enough to argue that individual humans do evolve. It is one of our strengths - we learn a lot as we grow up and become more "fit". Of course if you define evolution to require change in DNA that is not the case. There are also examples of organisms that undergo different development paths depending on their environment and in a certain sense they are also evolving.

I generally think it's important to try to understand what people mean and not worry so much about how they word things.

 

I apologize; I hadn't realized that you were using the broader sense of "evolution" meaning change of any kind.

When we are talking about natural selection in populations your examples are poor; a tadpole changing into a frog, or a person learning to use an abacus, do not relate to the selection of their children when they are removed from the picture - the frog's offspring will be born tadpoles, not frogs - the abacus user's child will not be born knowing how to use an abacus. Somatic adaptations - of any kind - are not passed on to the offspring.

Social adaptations can be passed on from one person to another, but the capacity to inherit social adaptations from another person (technology or whatever) is only dimly related to your genetic relation to them - you can learn how to unscrew a jar as easily from a gorilla as you can from your own parents.

Most people try to stick to genetic information as the medium for the passing down of adaptations, because fairly simple and predictable circumstances can easily wipe out the passing down of things like social adaptations - if the child is isolated from the parent, then none of their "evolution" is passed along except for whatever genetic adaptation was inherited by the child. No amount of evolution is meaningful in the sense of population adaptation if the gain is wiped out between generations.

 

I actually wasn't using evolution to mean change of any kind. My point was about language and context and how easy it would be for me to argue with you if I chose to misinterpret your words.

If you read the posts I think you'll agree that I wasn't talking about changes of any kind, that was my example of misinterpretation and how silly it is. I should have been more clear and labelled it that way.

Is there some reason you want to attack me? I come here to chat informally about interesting ideas and don't understand why you're being so antagonistic.

You seem like a smart enough fellow - why don't you contribute by putting forward interesting ideas instead of attacking the wording of other people's?

 

the model used in the Jernvall paper that can accurately predict tooth shapes actually throws all the inhibitors and activators on one heap. Therefore the question may arise if individual patterns of inhibitors and activators are that important. It is the combined effect that will eventually predict shape. This also explains why some activators can be exchanged for others in one system between 2 species. This is hardly a quantifiable effect in the sense you are aiming at ...probably. I don't know.

 

this might be in line with what you are looking for

 

I've read that paper ... my meta-terminology is a bit of a rip off from there actually. If you haven't read any of Uri Alon's papers I recommend them when it comes to evaluating genetic networks statistically. They haven't got the data to do direct cross species comparisons though.

Although they get at group of co-regulated genes, they know nothing about the structure of the regulatory networks generating them (ie gene A regulates targets 1 .. n). I think that's of primary importance. It is certainly a good first step and interesting work though.

 

I was looking at fig 1. I noticed that compared to flies and worms the human species only uses a small subset of genes for development Which seems a bit strange, because we don't have that much more genes than a fruitfly, and all these genes are conserved. In fact, vertebrates usually have more duplicated genes.

 

I recall from bio that certain base sequences are more vulnerable to mutation than others - that having to do with the chemical properties thereof.

It might be possible that the forms of these highly important cell regulatory genes tend to protect them from mutation, and so they have been passed along relatively unchanged because of their importance. The less protected forms of the cell regulators might have been weeded out by mutation.

 

Am I correct in thinking the Stuart extract says that genes can form interdependent pairs which can sometimes become structurally stable and can be transmitted across generations?

Or does he just say that within stable and 'fit' sets of genes pairs of genes tend to be structurally complimentary to each other?

Or did he say something else completely?

 

who is stuart?

 

Stuart Kim - the last author on that Science Article.

As a very morbid side-note, I had lunch with him when he visited my school and he told us that they had some C elegans on the Columbia. Apparently NASA doesn't report much on the negative health effects of being in space, but there are a number. So they were looking into whether it would be a good model organism and wanted to see simply if they would survive the trip. They did.

 

I should have said Stuart et al, authors of the extract you posted.

 

I'm going to try to read the Jernvall paper this weekend, but I think it's quite reasonable to group all activators together and all repressors together. However grouping activators and repressors together would be a bit odd.

As far as the oddity in number of human genes involved in development, I would imagine it's a reflection of the annotation. Human development isn't directly studied so even if human genes are involved in development and studied in human they would probably be clasified functionally, like "transcription" or "signaling".

From re-skimming the paper and my recollection, the point of the Stuart paper is that genes that are involved in the same pathway or protein complex need to be expressed together. So over evolution their correlation in expression is maintained, whereas some genes are coexpressed for more indirect reasons in a single organism. Therefore one gains information about genes with conserved co-expression to genes of known function. They never really comment on protein structure.

I hadn't noticed/recalled the first author's last name was Stuart ... sorry about the tangential morbid anecdote.

 

Bigbluehead,
Sorry if I was a bit testy ... was a bit stressed out regarding my thesis proposal defense yesterday. Clarity in communicating science is important and you were just enforcing the proscribed nomenclature. Not to mention the fact I probably wasn't that clear. My committee had some issues with my clarity yesterday themselves.

n a CpG (meaning the sequence CG rather than the CG basepair) the C is typically methylated, which increases the probability that a C will change to a T. This methylation is suppressed around transcription start sites which leads to a high local enrichment of CpG in these regions. This sequence is relatively rare everywhere else. So far as I know this is the only sequence specific mutation mechanism.

Generally, important genes mutate more slowly because of selection, not their specific sequence. Actually ubiquitously expressed genes, genes with many targets, and genes with many interaction partners all mutate more slowly because they have more constraints on their sequence, structure, and/or function. This is quite independent of CpG islands. If anything the large number of constraints might make instances of CpG in these genes more long lived not less, though I've not heard anything on this topic. Not really my area. Paul Samuel certainly would know more about that kind of thing.

 

Scilosopher! I spend most of my time here arguing social reform with and2000x and Galt; your "testiness" is like a gentle caress. I'm often too argumentative, and it was good of you to point it out.

The example (CpG) you mentioned here is the one I recalled from my Bio classes - I probably should have looked it up again before mentioning it, since it's been a little more than a year since I was in school.

So... now I have to clarify for myself... given the content of the Stuart paper, were you referring to meta-networks as mappings of relations of functional expression by looking at the related genes?

(I ask because I previously thought you were referring to mapping the functional expressions of an organism by picking them out individually. I had not realized you might be referring to the kind of relation mentioned in this paper, which would make for a quite different experiment...)