More Recent Comments

Tuesday, February 24, 2015

How do you explain the differences between chimpanzees. humans, and macaques?

Here's a figure from a paper by Marques-Bonet et al. (2009). It shows the differences between various human genomes (blue); between the human and chimp genomes (red); and between the human and macaque genomes (yellow).

The results are plotted a a fraction of sequence identity. (Convert to percent by multiplying by 100.) The window is 100 kbp (100,000 bp). Human chromosome 2 is on top and chromosome 7 is below.

Notice that the substitutions are pretty much randomly scattered over every part of the two chromosomes. The data is consistent with the idea that most of the DNA in those chromosomes is junk and most of the substitutions are nearly neutral mutations fixed by random genetic drift. The differences between each pair of species is consistent with an approximate molecular clock corresponding to a constant mutation rate over million of years. The absolute levels of sequence identity (i.e. 98-99% for chimp/human) is consistent with the time of divergence from a common ancestor based on the fossil record and other criteria.

Here are my questions. Is there any other explanation that accounts for the data? Is it possible to explain the results as adaptations—substitutions that are mostly fixed by natural selection? Is it possible to explain the results according to Intelligent Design Creationism?

I'm particularly interested in hearing from the creationists. What is your explanation?



Marques-Bonet, T., Ryder, O.A., and Eichler, E.E. (2009) Sequencing Primate Genomes: What Have We Learned? Ann. Rev. Genomics Hum. Genet. 10:355-386. [doi: 10.1146/annurev.genom.9.081307.164420]

83 comments :

Konrad said...

Easy - God inserted this signal into His creation as a puzzle to help Him distinguish the mediocre scientists from the great ones. His Grand Plan was figured out by Asimov back in 1980: http://en.wikipedia.org/wiki/The_Last_Answer

Mikkel Rumraket Rasmussen said...

"Is it possible to explain the results according to Intelligent Design Creationism? "

Yes, but only ad-hoc. It's what the designer wanted because the designer likes to play jokes on us!

Mikkel Rumraket Rasmussen said...

In any case, I would be particularly impressed if they could produce a design account that also explains mutational biases.

Joe Felsenstein said...

No creationists yet, so I'll guess what they will say.

Usually they focus on the percentage of identity between DNA of humans and chimps. They raise the issue of how you count insertions and deletions, and other chromosome rearrangements. Which is a nontrivial issue. They they quote somebody in their camp who has recalculated this and concluded that the divergence Is Much Bigger Than That.

Then they proclaim victory and disappear, without ever doing the same calculation for Human versus Macacque. Which would be Much Much Bigger Than That. But doing that calculation is a no-no because it would validate the inference that humans are much more closely related to chimps than we are to macacques.

Your graphs are an excellent example of the important fact that evolutionary trees inferred from different parts of the genome tend to confirm each other. It was that, plus fossils and similarity of developmental sequences, that persuaded most biologists in Darwin's era that common ancestry was reaL

Anonymous said...

Questions:

Do the blue dots reflect variation within the human population?

Also, the gaps in H, C and M genomes are because they're all up top close to 1 where there is no variation?

Wasn't CH 1 the fusion chromosome?

Unknown said...

You can already hear the excuses "same design, same designer," or they'll use that magic "information" that never seems to get calculated.

Mikkel Rumraket Rasmussen said...

Yes but why does the designer has this obsession with A/T nucleotides over G/C ? It's weird how it looks like the designer wants to insert the same kinds of mutations one expects on biochemical grounds alone.

John Harshman said...

Why does the designer like CpG doublets so much? I think its because he moves in mysterious ways his wonders to perform. But I'm not quite sure.

Alex SL said...

One assumes, however, that in Darwin's era "evolutionary trees inferred from different parts of the genome" meant something different that the DNA sequences of chromosome 2 and chromosome 7. But that is just the thing: the same works for skeletal structure and the anatomy of reproductive structures, and then there are fossils and biogeography to further support a tree-interpretation of biodiversity.

Bjørn Østman said...

"The data is consistent with the idea that most of the DNA in those chromosomes is junk and most of the substitutions are nearly neutral mutations fixed by random genetic drift."

Question: Is it also consistent with neutral mutations fixed by genetic draft/hitchhiking?

Joe Felsenstein said...

Oops, you're right of course, no genome available then. Different parts of the anatomy, I should have said.

Joe Felsenstein said...

In short, yes. Both yield the same expected difference.

Larry Moran said...

In short, no. One the reasons I posted this figure is to get you (and Joe) to post an adaptationist explanation. (I saw your question on Twitter.)

Assiming that I understand what you mean by "genetic draft/hitchhiking" then I think the data shown here is NOT consistent without special pleading.

What you are talking about is purifying selection where a part of the genome is swept to fixation because of the presence of a fairly strong beneficial allele near the middle. Presumably this has to happen quite frequently in order to cover the entire chromosome in just a few million years.

Such a mechanism is unlikely to generate a fairly even distribution of substitutions across the entire length of the chromosome but that's what we see.

Also, I don't think that the mechanism you are suggesting is consistent with the observation that the overall rate of fixation of these alleles is close to the mutation rate. I could be wrong about that since Joe seems to think it does.

If it's true that only 10% of the genome is functional then all of those adaptive mutations have to be confined to that small percentage of the genome and the other 90% is "drafted" to fixation. Can either of you do a brief calculation of the number of beneficial mutations that have to be swept to fixation in order to generate the three patterns we see in the figure?

peer said...

Question: The conclusion is based on n=1 data or a bigger dataset including all polymorphisms?

Then, I see that Moran says that ... "the data is consistent with the idea that most of the DNA in those chromosomes is junk..."

I think this is unwarrented, rather the conclusion should be that there is no-selection on these sequences (and thus fully consistent with the frontloading hypothesis of some ID proponents).

The major part of genes resides in the genome without selection, operating in cooperative networks (as might be expected from an ID point of view).

Furthermore, Moran says the data are "consistent with the time of divergence from a common ancestor based on the fossil record...".

My question: what common ancestor from the fossil record do you have in mind?

Marcoli said...

My guess is that the blue dots at and near 1.0 are human variations. The gaps in the C and M aligned chromosomes should be the centromeres. Those sequences are probably conserved.
But I have not looked at the paper.

peer said...

With beneficial mutations you mean an addition of novel genes? HARs? Through acquisition of novel gene regulatory elements (TEs) that drive de novo previous non-coding/non-functional DNA sequences, which then instanteniously adopt novel functions with a highly selectable value? Is that what you mean?

If so, welcome to the world of frontloading.

pliny said...

The fusion chromosome in humans is #2.

Joe Felsenstein said...

Larry, genetic draft is when mutant alleles, particularly neutral ones, are pulled up or down in gene frequency by nearby mutations that are under selection. The result is that those neutral alleles make large gene frequency changes up and down. The upshot is that it has the same effect on them as locally enhanced genetic drift.

The effect of genetic draft on distributions of gene frequencies, and of haplotypes, is not the same as genetic drift. But to see that you would need to have reasonable size samples of haplotypes in each species.

These data are, in effect, single sequences, one from each species. We know that for single sequences the expectations of divergence are basically the same for genetic draft as for genetic drift. One haplotype sweeps through the population, carrying closely linked alleles to fixation, or nearly so. The effect is that our prediction of divergence between those single sequences is still the same. (There are some differences of prediction but they are that hitchhikling events in the common ancestor will reduce the variation in divergence from one region in the genome to another).

Larry, genetic draft still means most of the sites you see in the above graphs are neutral, The prediction is nearly the same for single genomes, one from each species. So I still say that (to a first approximation) the predicted divergence between species is unaffected by whether there are some selective sweeps.

Petrushka said...

Why would you expect to find an ancestor in the fossil record?

peer said...

I do not expect anything from the fossil record. I wondered why Moran claimed these data to be consistent with the fossil record.



Larry Moran said...

@Peer Terborg

I'm anxiously awaiting your detailed explanation of the data. When can I expect you to post?

Larry Moran said...

Joe,

I think I understand what you are describing but it does seem a little counter-intuitive. I'm pretty sure your conclusions require a lot of recombination but that's okay because there IS a lot of recombination.

Imagine that there was no recombination. In that case the entire chromosome would be swept to fixation as long as there was a single beneficial allele with a fairly strong selection coefficient. (This would depend on the population size, of course.)

If you compared the sequence of that chromosome right after fixation to the sequence of the orthologous chromosome in a closely related species it would look like rate of fixation of all the alleles was much faster than the rate predicted if they were being fixed by random genetic drift, right?

Under those circumstances, you would know that something strange is going on, right?

I assume that this effect gets diluted over time so that after a million years or so the difference in rate wouldn't be noticeable. I also assume that the effect will be less pronounced when recombination is common because the sweep will be restricted to just a small part of the chromosome. Am I right?

I do understand that the occasional selective sweep will be difficult to detect at this level of resolution, especially if it occurred more than one hundred thousand years ago.

Is that all that you and Bjørn are saying? Or, are you both trying to make the case that adaptation is much more common that most people realize? Or are you just challenging the conclusion that most of the neutral alleles are fixed by random genetic drift because adaptive sweeps are very common?

P.S. I get the point that you can't tell for sure if a given allele is fixed in the population based on a single sequence. Isn't it okay to make the assumption that this is true for most of the differences between the macaque and human genomes?

Bjørn Østman said...

Has such a comparison between drift and draft ben done? I think I could do the calculation/simulation fairly easily, but I am unsure what data to extract for the comparison (and I guess what parameter values to use (pop-size, mut rate, genome size, recombination rate).

Larry, I wonder about both of those questions: is adaptation in humans much more common than we think and are adaptive sweeps common enough for neutral alleles to sweep by draft. From experimental evolution (LTEE and the like) the conclusion seems to be that there's always some beneficial mutations sweeping, and that's what got me wondering about other types of organisms.

Unknown said...

I see an issue with language here and it applies to both sides of this argument. An allele is neutral iff s=0 and near neutral iff s~0. And s records a correlation between fitness and the presence of the allele in question. So when a mutation is associated with higher fitness because there is another mutation nearby that affects fitness, while not affecting fitness itself that mutation is not neutral. In the case Larry gives - no recombination along chromosomes and no novel mutations - there is no variance in the selection coefficients among loci on any particular chromosome.
I think there is an issue with calling an allele neutral, but meaning that it would be neutral given independent inheritance. We can make such a distinction when we are not talking about neutrality in that an allele is beneficial is s>0 and adaptive if s would be greater than 0 given linkage equilibrium. I think it would be helpful to have a similar terminological distinction for s=0 and s=0|LE.

Bjørn Østman said...

I think the language is quite clear. If a mutation is neutral on its own, then it is still neutral even if linked with a nearby (in time or space) beneficial mutation. We are talking about the functional/phenotypic effect that the mutation has. Why would you want to confuse things by not calling it neutral?

(Also, a mutation can be effectively neutral if it is less than ~1/N.)

Mikkel Rumraket Rasmussen said...

"I think this is unwarrented, rather the conclusion should be that there is no-selection on these sequences (and thus fully consistent with the frontloading hypothesis of some ID proponents)."

The "frontloading" retort is entirely ad-hoc, it doesn't really follow from any principle of ID.

If front loaded is erected once, some time close to the origin of life, then It doesn't explain ORFan sequences and so has to be artificially erected again ad-hoc for every ORFan sequence unique to some species or lineage.

There's no underlying logical framework that says how much front loading there should be, how often or when in particular the designer intervenes to "front-load" some ancestral species with new genetic material, or anything at all. No details are really explained, no quantifiable predictions are made that can be tested statistically. It's all after-the-fact rationalization.

The idea also seems to allow for universal common descent and almost all of molecular evolution to just proceed by known natural mechanisms of speciation through isolation and subsequent mutational divergence.

Joe Felsenstein said...

@Simon: I disagree strongly with your definition of neutral. If genotypes AA, Aa, and aa at locus A all have equal fitnesses, then these two alleles are neutral (with respect to each other). If there is a locus B nearby whose genotypes differ in fitness, then I would say the variants at locus A are still neutral, whether or not B is in linkage disequilibrium with A. Sure, A may change frequency as if it is selected, but it's still neutral in my book (and as I am teaching out of that book right now, I need to have a consistent definition -- I'm teaching the neutral theory to them 20 minutes from now, as it happens).

Joe Felsenstein said...

Sorry, my link to the book should have been this instead.

Unknown said...

I think your post shows precisely how the language is not clear. In general we are looking at two different causal processes:
a) How the fitness of individuals affects allele frequencies
b) How various alleles affect individuals fitness
You can build simplified models where the difference doesn't matter, mainly by excluding linkage. But if you want to have a model that includes linkage, you need to differentiate between the two.
Now in your post you make two claims. One is that the neutrality of an allele depends on the "functional/phenotypic effect". That refers to (b) above. You also state that a mutation can be effectively neutral if |s|<1/N. That is a statement about (a).

However since a and b are different, you encounter cases where the two do not match. In particular you can have cases where there is no phenotypic effect (neutral according to your first statement), but through linkage s>50/N (not anywhere near neutral according to your second statement). By using the same term defined through two different processes, you end up with a situation where you are not clear.
As noted before we do have this distinction with beneficial and adaptive mutations, where beneficial refers to (a) and adaptive to (b). We need terms to deal with neutral (a) and neutral (b).

Paul McBride said...

Larry says: "If you compared the sequence of that chromosome right after fixation to the sequence of the orthologous chromosome in a closely related species it would look like rate of fixation of all the alleles was much faster than the rate predicted if they were being fixed by random genetic drift, right? "

The rate of fixation is higher, but not the substitution rate, and it is usually the substitution rate that is estimated. The lack of difference is because the terminal branches on phylogenetic trees from which substitution rates are estimated contain both fixed and polymorphic differences.

Consider the individual in which that single beneficial mutation arises. It has inherited all of the neutral polymorphisms that are soon going to fix under the selective sweep. If you sequenced the chromosome in question in it and another individual from the population and compared them to a closely related species neither comparison would reveal anything surprising, as the other individual should have an approximately equal number of polymorphisms and many shared ones (given an ideal population). When the selective sweep happens, the rest of the variation at that chromosome in the population is wiped out, so diversity is heavily reduced, but other than that nothing untoward or special happens.

In fact, the selective sweep does not even change the odds of a given neutral mutation fixing. The chance of any given neutral mutation being present in the genome where the beneficial mutation arises and hence hitchhiking to fixation is equal to its frequency in the population, and its frequency in the population is also its probability of fixing in the absence of selection. This result is nicely demonstrate by Birky and Walsh 1988.

Unknown said...

My post above was directed at Bjørn, so I need to adress Joes post. I find it hard to makes sense of it, mainly because it's not entirely clear what he means when he states that locus a is neutral. I.e. the same issue again, because if we restrict ourselves to the haploid case, we can't have all of the following:
a) a and A individuals have the same mean fitness
b) ab and aB individuals have the same fitness and Ab and AB individuals have the same fitness
c) locus B is not neutral
d) there is LD between the two loci

At least one of the 4 statements above is false and my guess is that Joe would not include (a). But it's a guess precisely because neutral is used when (b) is false regularly. I agree that we need to use the term consistently, hence my call for clarifying language.

Joe Felsenstein said...

Larry: what Paul McBride said. A little more explanation. Suppose that we compare a human chromosome to the corresponding one in macacque (assuming there is one). We pick a random copy from macacque and get 94% sequence identity. We'd get about the same result from other copies sampled from macacque.

Now suppose that there was a selective sweep in macacque that takes one of those copies and raises it to 100% frequency. Does that affect what sequence identity we see when we sample one copy from macacque? No.

It does affect the result we get when we sample two copies from the macacque population, certainly. But it does not affect the average sequence identity between human and macacque.

Joe Felsenstein said...

Simon: There are two many "not"s and "false"s in your example, I get mixed up trying to answer the questions you posed.

I call a change of allele A to allele a neutral if it makes no difference in fitness no matter which genotype it occurs in.

In your (haploid) example, A and a may not have the same mean fitness. But as long as the fitness of AB is the same as the fitness of aB, and the fitness of Ab is the same as the fitness of ab, then I call the change from A to a neutral.

I think a lot of other people would agree with me, not with you. So there.

Unknown said...

I don't even disagree with you, I just want this to be clear. By your definition above we do not know the probability of fixation of a novel neutral allele, because it depends on factors beyond its neutrality. Nor would we necessarily expect clocklike divergence for neutral mutations.
Now, Larrys OP explicitly links neutrality to clocklike divergence (and using the alternative of s=0 it's clear that given constant µ, you obtain clocklike divergence from neutrality in this sense). That suggests you might be using different concepts of neutrality and hence talking past one another to some degree.

Konrad said...

I can think of one regular commenter who might say something like that, but punctuated a little differently:

"Usually they focus.

On the percentage of identity between DNA. Humans and chimps? They raise the issue.

How you count insertions and deletions? and other chromosome rearrangements? Which is a nontrivial issue.

They they quote somebody in their camp. Who has recalculated this? And concluded.

The divergence Is Much Bigger Than That."

The whole truth said...

I don't see how asserting a larger percentage of differences between human and chimp DNA (or any other differences) helps the creationists' claims about 'Special Creation' of humans unless there were NO similarities (matches) between chimps and humans (in DNA or otherwise) and NO similarities between humans and any other biological entities (in DNA or otherwise). If humans are unique, exceptional, 'Specially Created in the image of God', and are not related to any other biological entities, then why are there any similarities (in DNA or otherwise) between humans and other biological entities and why are there so many similarities between humans and other primates, and especially other apes?

Another thing to consider is the dependencies and other interactions humans have on and/or with other biological entities (e.g. gut microbiota, viruses, parasites, and the biological entities that humans eat).

Yeah, I know, creationists conjure up lots of BS about why there are any similarities between humans and chimps (or humans and other biological entities) and about dependencies and other interactions, but none of their BS assertions support their claims about 'Special Creation' of humans.

socol said...

I have a question:
Does the fact that the molecular clock based on neutral mutations obviously would lead to understand that would be consistent with species differences . Because these sequence differences are mostly large neutral mutations . ?
But I understand that these neutral substitutions,
not generate evolutionary changes that support the evolutionary interpretation of the fossil record .
Do not we would need a molecular clock based on mutations subjected to selection to ensure it is concluded here ?

AllanMiller said...

As soon as there is selection you get into rate difficulties because it is affected by population size and the strength of the selection itself. You need a clock that ticks constantly (as an ideal, though this is not necessarily practical), such as a neutral clock. Most fossil differences, meanwhile, are likely to be due to just a few alleles, and phenotypic change could readily be 'mostly adaptive' without any conflict with the observation that most molecular change is neutral.

peer said...

The general theory of evolution (i.e. universal common descent) was falsified a long time ago by biological observations summarized by the late professor John A Davsion in his evolutionary manifesto.

A special theory of evolution (limited common descent, more than one LUCA) is still valid, though, and in accordance with fronloading theories.

Currently, in the absence of a scintilla of evidence for abiogenesis and loads of evidence against it, there is no better scientific explanation than frontloading.

Peer

peer said...

Allan, your statements are based on what observations? What is your evidece for your statement that fossil differences are due to a few alleles?

Or do you mean karyotypic mutations (transocations, inversions, duplications of parts of or whole chromosomes) that effect the genome structure and alter gene regulatory patterns as a whole (and which is a major driving force for speciation within the frontloading theory).

In my opinion, your alleles are too NeoDarwinian/population genetics. It will never bring anything novel. Karyotypic changes may release novel cryptic information, however, previously not expressed bauplane.

peer said...

Prof Moran,

Your questions:

>Is it possible to explain the results as adaptations—substitutions that are mostly fixed by natural selection?<

I agree with you that they are not for reasons Haldane calculated about 70 years ago.

>Is it possible to explain the results according to Intelligent Design Creationism?<

In my opionion, these results can be brought into accordance with a frontloading theory as publsihed in a serie of articles in the J of Creation 2008/2009.

Mikkel Rumraket Rasmussen said...

"The general theory of evolution (i.e. universal common descent) was falsified a long time ago by biological observations summarized by the late professor John A Davsion in his evolutionary manifesto.

A special theory of evolution (limited common descent, more than one LUCA) is still valid, though, and in accordance with fronloading theories. "


It's funny because this view has apparently failed to catch on. When you're forced to make shit up to defend doctrine it doesn't look good.

"Currently, in the absence of a scintilla of evidence for abiogenesis"

What does that even mean, what evidence should there be? It happened over 3.5 billion years ago. Classic argument from ignorance, that fact that we don't know how it happened isn't evidence that it didn't happen.

"and loads of evidence against it"

You have evidence against abiogenesis? What you probably mean is you have arguments against specific scenarios, right?

Larry Moran said...

@Peer Terborg

The differences between the natives of Japan and the pygmies of central Africa are at least as great as the differences between most closely related species.

Do you think the differences within our species are due to karyotype changes or to just a few allele differences?

AllanMiller said...

@Peer Terborg,

What is your evidece for your statement that fossil differences are due to a few alleles?

Most differences observed in fossils will likely be due to regulatory changes in development, which represents rather a small subset of the functional genome, itself a small subset of the overall genome. Most differences between genomes will be in the other bit. Even if every functional part has changed, that's still much smaller than the rest.

AllanMiller said...

Perhaps, just to clarify, I should have said visible differences were due to 'a small proportion' rather than 'few'.

Joe Felsenstein said...

Simon: You said By your definition above we do not know the probability of fixation of a novel neutral allele, because it depends on factors beyond its neutrality. Nor would we necessarily expect clocklike divergence for neutral mutations.

If we average over all the copies of all haplotypes in which a novel neutral allele can arise, the probability of fixation of it or its descendants is still 1/(2N), And we do expect clocklike divergence, once we take into account the particular coalescent trees ("gene trees") that hold for each region of the genome.

Unknown said...

What this boils down to is "we can ignore the effect of linkage disequilibrium, if we assume that linkage disequilibrium has no effect". Which is a pretty trivial result, I think. Because if we do a LLN for the appearance of a novel allele, we of course end up with no linkage.
Which BTW, is now keeping statements a-c and rejecting d, as opposed to your previous post in which you dropped a. You can not have your cake and eat it, too.

Joe Felsenstein said...

Simon: We're still in disagreement. I *am* still dropping (a), and and am defining neutrality as AB and aB having the same fitness, and also aB and ab having the same fitness (if these are the only two loci segregating). I am not dropping LD.

More precisely, my definition of neutrality is that the effect of mutating A into a must be no change in fitness, no matter what haplotype we do it in. Note that this is *not* a statement about marginal fitnesses across all possible distributions of haplotypes. It is a statement about the fitness differences between two haplotypes, ones that differ only at the A locus.
If we have diploid organisms, we can make the analogous definition of neutrality involving diploid genotypes.

With neutrality defined this way, if all of the haplotypes have the A allele, and we pick a random haplotype copy and mutate the A to a, I claim it is true that the probability of ultimate fixation of a (or neutral alleles descended from that copy) is in fact 1/(2N) irrespective of what fancy LD and selection is going on. At least it is if all 2N copies at the A locus have a coalescent.

Unknown said...

I can see what you mean, assuming B type individuals have a probability of going to fixation of PB and a frequency of pB, then given that we randomly pick an individual to mutate to a, we get
Pa=pB*PB*1/(NpB)+(1-pB)*(1-PB)*1/(N(1-pB))=PB/N+(1-PB)/N=1/N

However this does have the effect of removing linkage from consideration, as we look at a sample space where we could both have linkage with B and b and the two cancel out in our assessment. I.e. you are calculating the probability that a mutation to a would be fixed assuming it will occur in one random individual, which differs from the probability that a mutation to a will be fixed, given that it has occurred in an individual that is either B or b, which are
Pa|B type=PB/(NpB)
and
Pa|b type=(1-PB)/(N(1-pB))
Which are only equal to 1/N, if PB=pB and thus locus B is neutral.
I stand by my original claim, averaging over all individuals where the mutation could occur is equivalent to removing the effects of LD from the model.

socol said...

We have a "neutral " but by itself does not tell me anything, we need to measure at the clock when those mutations that generate an evolutionary divergence (which is difficult ) occur clock. And thus determine the evolutionary times and then compare the inferred the fossil record .
What I see is that they are taking time data obtained by the dating of fossils and placing on the molecular clock. Do you really are getting some consistency between what obtained with the molecular clock and fossils . ?

Joe Felsenstein said...

Simon: Fine. You stand by your statement that what I have done is "equivalent to removing the effects of LD from the model". I will stand by my statement that if we mutate the A locus in a random (haploid) individual, that the probability of fixation of it-or-its-descendants is 1/N no matter what pattern of LD we have. Your nice proof shows this. We can both be right.

But let's remember what the original question was: would the divergence between human and macacque be affected by selective sweeps at some sites? The answer is yes, for those sites, but no, for any neutral sites, even ones closely linked to the selected sites. They mutate in randomly chosen haplotypes. So overall, the answer is no. The rate of change is unaffected by the selective sweeps, since almost all of the sites we see in the graph in the original post are the neutral ones, and we do have an approximate molecular clock.

Unknown said...

I think we are mostly in agreement now. The one difference I see is that you state that there is no difference for any neutral sites, while I'd say that there is one for quite a few of them, but the effect across many sites averages out. So you would have neutral sites with substitution rates above the neutral rate, neutral sites with substitution rates below the neutral rate (depending on whether mutant alleles were swept to fixation or extinction be nearby sites under positive or negative selection), but the average substitution rate is equal to the neutral rate, giving us a MC.

NickM said...

Davison liiiiiiiiivesssss.....aaaaaagggggaaaaaiiiiiinnnnn!

Jon Fleming said...

I love it so!


sorry I had to.

Diogenes said...

Terborg: In my opionion, these results can be brought into accordance with a frontloading theory as publsihed in a serie of articles in the J of Creation 2008/2009.

Citing the well-respected Journal of Creation. Low, even by creationist standards.

The general theory of evolution (i.e. universal common descent) was falsified a long time ago by biological observations summarized by the late professor John A Davsion in his evolutionary manifesto.

Oh, what bullshit. We've seen every anti-evolution argument so we know this is BS.

1. Front-loading is 100% pseudoscience. It does not logically follow from the claimed hypothesis. I define pseudoscience based on tricks such as, predictions that don't follow from the claimed hypothesis, and the use of equivocation to evade falsification.

It's BS to claim:

1. Hypothesis: living things are intelligently designed, therefore

2. Therefore, Ambulocetus was front-loaded with a program to evolve into a blue whale, Chinese river dolphin etc, Australopithecus was front-loaded with a program to evolve into a Zinjanthropus, Neanderthal, human, etc., Psittacosaurus was front-loaded with a program to evolve into a Triceratops, Chasmosaurus, etc., moeritherius was front-loaed with a program to evolve into Ambelodon, gomphotheres, Mastodon, mammoth, Asian elephant, African elephant, etc. etc. etc. etc. etc. etc. etc. ad infinitum.

No, that doesn't logically follow and it's pseudoscience.

2. Front-loading is just orthogenetic evolution warmed over. Orthogenesis falsified by the 1950's.

3. Just an attempt to evade the fact that we have observed evolution creating specified complexity and increases in information. Therefore Intelligent Design is DEAD and this is just an attempt to evade falsification.

You're not worth our time, but for fun we'll make you dance and look like an idiot by asking again:

Do you, Terborg, really believe that Ambulocetus was front-loaded with a program to evolve into a blue whale, Chinese river dolphin etc, Australopithecus was front-loaded with a program to evolve into a Zinjanthropus, Neanderthal, human, etc., Psittacosaurus was front-loaded with a program to evolve into a Triceratops, Chasmosaurus, etc., moeritherius was front-loaed with a program to evolve into Ambelodon, gomphotheres, Mastodon, mammoth, Asian elephant, African elephant, etc. etc. etc. etc. etc. etc. etc. ad infinitum? Do you have a stick of evidence for that?

Diogenes said...

"Karyotypic changes may release novel cryptic information, however, previously not expressed bauplane."

IDiot just said "cryptic information."

Please analyze the genome of any species anywhere and identify > 5 base pairs containing "cryptic information", and measure how many bits of "cryptic information" there is in them.

Oh wait, creationists can't.

Oh $%^&, how about if you just quantify NON-cryptic information. Take any genetic sequence, I don't know, GULOP or something. How many bits of NON-cryptic information does it contain? How many $%^&ing bits?

By what equation would you measure NON-cryptic information? Shouldn't you $%^&ing IDiots cough up an equation for NON-cryptic information before you start blithering about a CRYPTIC information that is invisible and you can neither see NOR define?

Wouldn't it be nice if you $%^&ing IDiots could define or measure NON-cryptic information before you began invoking "cryptic information" to evade falsification of your BS claims?

Jibbers Crabst, we observed evolution producing increases in complexity, functional information, specified complexity, irreducible complexity! ID was falsified every way it could be falsified! You lose, that's all!

Let me put it this way. You're cripples and most pro hockey players don't want to play hockey against cripples. It wouldn't be sporting, see. But then you're assholes and you keep accusing pro hockey players of being in a conspiracy against you. So some of us got mad and went out on the ice with you cripples and scored goal after goal after goal and made you look like boobs. Now the scoreboard clearly shows we got all the goals and you got none!!!

But your response is to rub your hands together and say, "Ah yess, but there is a CRYPTIC SCOREBOARD that is invisible and no one can see it, yessss....."

peer said...

@ Moran

Moran: The differences between the natives of Japan and the pygmies of central Africa are at least as great as the differences between most closely related species.

Yep, and it s even more than Neodarwinian allelle-worshippers think it is. Most variation resides in transposable and transposed genetic elements, non-coding regulatory sequences, etc. It is not the genes, but the expression of genes that makes the phenotypes.

Moran: Do you think the differences within our species are due to karyotype changes or to just a few allele differences?

No, but some may. Duplications and deletions are easily exerted by repetitive sequences. The two genome projects demonstrated an almost 10% content difference. Alleles? No, duplications and deletions.

The point is that you will hardy find real intra-human whole genome comparisons in the literature because repetitions are so hard to determine using NGS methods. I was involved in a project to determine TEs (transposable and transposed elements, which are in fact nothing but juming promoters and enhancers) in two populations of humans but we were unsuccessful because of this shortcoming.

peer said...

@Diogenes,

Starting with namecalling and swearing is very weak, D.

What your response on information demonstrates is that you are not a biologist, not even a scientist would be my guess. If you were, you would have known that we now know of several genes that have cryptic promotors, which drive cryptic genetic information in response to alternative, more error prone RNA polymerases. If you want to discuss modern biology better take some advanced courses.

Abou definitions> it would be nice if the term evolution is defined before discussion takes place. Not even Moran definied it, here, as far as I can see. Now wonder this evo-crea debate is endless. Because you guys do not define.

peer said...

@ Diogenes. Jibbers Crabst, we observed evolution producing increases in complexity, functional information, specified complexity, irreducible complexity! ID was falsified every way it could be falsified! You lose, that's all!

It is not about winning or losing. It is about science. If your claims were true for biology, everybody would know about it. And there would be not be a scientific ID community.

So, Diogenes, leave your tub, and please expand with a scientific observation or reference which can be checked ad reproduced in the lab.

AllanMiller said...

Why can't duplications and insertions be alleles?

peer said...

@ Allan,

Most differences observed in fossils will likely be due to regulatory changes in development, which represents rather a small subset of the functional genome, itself a small subset of the overall genome. Most differences between genomes will be in the other bit. Even if every functional part has changed, that's still much smaller than the rest.

So, no alleles...but regulatory changes. TEs, suffling of preexisting genetic info...No addition of new genes required....frontloading.

I already thought so, because the neodarwinian concept of biology is completely down and out.

peer said...

So, now you have to define what an allele is. Say, the entire coding part of a gene duplicates and now it is twofold present in the genome...do we call the entire region an allele. If so, than you are right. But in science we should be more specific. Using the same term for duplication and insertion for allele makes it very confusing. That-s why we call them duplications and deletions. Allele is a very non=specific term.

AllanMiller said...

socol,

Do you really are getting some consistency between what obtained with the molecular clock and fossils . ?

Within the known limitations of the method, yes. You'll really have to be a bit more specific about the period you're concerned about. Fossils are used to calibrate molecular data, sure - there are numerous mechanisms that may cause the clock to drift from a steady tick. But tick it does. It records relative time well enough, but on absolute time it's not as steady as radio-isotopes, which is why fossils (where available) tend to be the overriding reference. But there is a lag here, between the point of divergence from a common ancestor and the appearance of forms in fossils, and so molecular clocks would tend to point deeper than fossils even in an ideal world.

peer said...

@ Moran,

Moran: The differences between the natives of Japan and the pygmies of central Africa are at least as great as the differences between most closely related species.

After rereading this statement> what difference do zou refer to...SNPs:?

As far as I know nobody ever compared whole genomes. I am pretty sure that if one perfoms that unfiltered we will see a lot of surprises, which will blow NeoDarwinism ideology further out of the water.

AllanMiller said...

Not really. An allele is a variant of a particular genetic sequence of interest. If you have a region ->xxxyyy<-, where the arrows represent positionally identifiable flanking sequence, then ->xxxxyy<-, ->yyy<-, ->xxxxxxyyy<-, -><- etc are all alleles of that sequence.

I think confusion often arises because people think in terms of functional regions - protein coding, for example, so a duplication of a protein coding gene might seem like '2 alleles'. But the 2-copy genomic region is an allele of the 1-copy genomic region, within the population. As far as a diploid is concerned, alleles are just regions that can be homozygous or heterozygous within it. The boundaries of the allelic region can be narrowed or widened. So ->GeneGENE<- can be seen as heterozygous with ->Gene<-, or as two separate allelic regions, the homozygous ->Gene<- and heterozygous ->GENE<- and -><-. A gap can be an allele of a non-gap.

peer said...

The whole NGS tecnology to analyse and compare genomes is so immensely NeoDarwinian biased, pretending there is nothing but protein/coding exones, we missed a lot!

A few years ago I received a grant to study and compare the TE compartments of the genome. To mzysurprise not a single NGS platform had factored in that the difference between species and health and disease could be in the dark regions of the genome! Incredible isnt it? They were unable to handle it and my colleagues and I had to invent our own methods. Neodarwinian bias is now slowly being removed by many of my friends of ENCODE. And that is the way to go.

AllanMiller said...

I was involved in a project to determine TEs (transposable and transposed elements, which are in fact nothing but jum[p]ing promoters and enhancers) [...]

Although some TEs can end up serving such a functional role, it's a bit of a stretch to say that's what they are. They also donate coding sequence, and serve to disrupt equal crossover, and intron excision, and cause fatal mutation. None of these can be seen as, especially, a role.

AllanMiller said...

PT: So, no alleles...but regulatory changes. TEs, suffling of preexisting genetic info...No addition of new genes required....frontloading.

That's something of a rabbit out a hat, there. "stuff ... more stuff ... hey presto! frontloading!". You think that there are no genetic changes beyond 'shuffling'? And what is it with you people and ellipses?

AllanMiller said...

pretending there is nothing but protein/coding exones

I think your implication that 'neo-Darwinians' argue or believe that there is nothing to evolution (descent with heritable modification) beyond changes to coding regions is something of a strawman.

The whole truth said...

Peer, have the "front loading hypothesis" or "front loading theories" that you are promoting been checked and reproduced in the lab?

What are the details of the "front loading hypothesis" or "front loading theories" that you accept as being accurate?

Has "A special theory of evolution (limited common descent, more than one LUCA)" been checked and reproduced in the lab?

Do you believe in 'special creation' of humans from dust and a rib? If so, has 'special creation' of humans from dust and a rib been checked and reproduced in the lab?

Do you believe and claim that yhwh-jesus-holy-ghost exist and that they are 'God'? If so, has the existence of yhwh-jesus-holy-ghost been checked and reproduced in the lab, and has the veracity of your claim that yhwh-jesus-holy-ghost are 'God' been checked and reproduced in the lab?

Larry Moran said...

Neodarwinian bias is now slowly being removed by many of my friends of ENCODE. And that is the way to go.

ROTFLMFHO!!!!!!

Mikkel Rumraket Rasmussen said...

"Neodarwinian bias is now slowly being removed by many of my friends of ENCODE."

Holy fucking shit, did he actually write this?

Anonymous said...

PT,

So, no alleles...but regulatory changes. TEs, suffling of preexisting genetic info...No addition of new genes required....frontloading.

Alleles are different versions of a locus in a population. So a few regulatory differences would still be differences in a few alleles. Even if it wasn't, even if the definitions didn't match, the important point Allan was making is that few changes have phenotypic effects, with "few" being relative to the rest of changes, which would be mostly neutral or nearly neutral. You missed the point in exchange for a meaningless diatribe, attempting to make your case on definitions, rather than on understanding the point: that among many changes, just a relatively few might account for phenotypic/visible differences.

Try some better understanding next time.

Anonymous said...

They also donate coding sequence, and serve to disrupt equal crossover, and intron excision, and cause fatal mutation. None of these can be seen as, especially, a role.

Of course they can! TEs are in fact nothing but jum[p]ing mutators, mental disease producers, retropseudogene insertors, copy-number-variation enhancers, etc, etc, etc. They are all of that and more...frontloading!

Faizal Ali said...

If your claims were true for biology, everybody would know about it. And there would be not be a scientific ID community.

Just quoting for posterity. You have to almost stand in awe at the stupidity of this argument.

The whole truth said...

Yahoo posted an article today about humans/chimps:

http://news.yahoo.com/big-brain-gene-found-humans-not-chimps-192932726.html

socol said...

Alan
I think we agree is that the molecular clock itself is not reliable, and reading depends heavily on what was found with the dating by radioisotope.

Then returning to the post of MORAN, says:

"The Differences between each pair of species is consistent With
an approximate molecular clock Corresponding to a constant mutation
rate over million of years.The absolute levels of sequence identity
(Ie 98-99% for chimp / human) is consistent With The time of
divergence from a common ancestor based on the fossil record and
other criteria. "

But if the molecular clock is calibrated dating
fossils.
Then implicitly when measured with the clock differences between species; the results will HAVE to be consistent with what is known about divergence times (Fossil) between humans and chimpanzees or humans and macaques.
Thus, the supposed consistency really is unclear or may not exist.

The molecular clock just tells me there is a mutation from time to time.
But these substitutions do not imply that evolutionary change is occurring (selectively positive mutation or negative) is like we have a burning car but neutral; the motor (clock) is revolutionizing but the car does not move (evolutionary change) and we can not determine at what point does it but only with reference to an external agent (fossil).
So if we take the differences between pairs of species genomes and measure with a molecular clock have to do some delicate assumptions:

For example; 1% of Sequence identy differentiates human chimp genome is unique for the first (but gorilla genome have identical regions corresponding to the 1% that man would suppose exclusive ) .

Also assume that the differentiating sequence was given at that precise combination within a universe of probabilities.
Considering the above , I consider that there is no theoretical basis for comparative claims between the results obtained with a molecular clock and dated with the fossil record , because in its definition are engaged.

AllanMiller said...

In the case of a simple bifurcation to 2 modern species, there is no issue. You count the sequence differences between them, and you measure the rate at which the modern clock ticks. In a simplistic model, basic arithmetic gives you a rough time in years since divergence. No fossils are involved. Say you get 10 million years by this method.

Then say you find a fossil which is close to the main branch of both your modern species. You date it, get 7 million years, and consider that date to be more reliable. So the molecular clock is out. But it's still clearly functioned as a clock, it wasn't defined circularly or by fossil calibration, and it wasn't a million miles out. You take the recalibration into account when you go further back, using your clade to count differences with a sister group.

Of course, you can get a lot more sophisticated and account for substitution biases, concentrate on definitively silent sites, and so on, to improve the accuracy of the clock. But I don't see the issue. When you date that oldest fossil, you have an individual prior to the emergence of the phenotypic differences that seem to concern you. Fossils on either branch, harbouring some of the incipient differences between the modern species, are no use for calibrating the molecular clock, which starts with molecular data from 2 extant leaves and attempts to date the node connecting them.

AllanMiller said...

Correction - intervening fossils do have a use, since the divergence clock cannot be younger than them.

Joe Felsenstein said...

I notice that "Terborg" has not answered. Apparently Terborg was aware of what would happen if he tried to answer.

The whole "front loading" hypothesis doesn't make sense for two reasons:

(1) It is incoherent, as Diogenes implied. Where is the "front"? 500 million years ago? Was the genetic program to make, say, whales already built into Cambrian chordates? Along with the program to make puffer fish?

(2) It could not be maintained in the face of mutation. If the whale program was sitting there in the genome 500 million years ago, what prevented it from being eroded by mutation before it was time to come up with whales?

A front-loading designer not only has to build in the structures of all possible descendants, but He has to make sure that the programs are maintained with no disruptive mutations occurring before the program is needed. And each time there is a speciation the programs somehow have to know that in one lineage the whale program won't be needed but in the other it will.

The Designer is awfully busy in that "scientific" view: not only is it true that "His eye is on the sparrow" but He's busy preventing the wrong kinds of mutations in all species, throughout the whole process of front-loaded evolution.

In which case it is not clear why He needed to front-load in the first place, if He could just order up the right mutations when needed.

socol said...

Thanks for your interest in answering my questions. There are points given for more discussion ... but thanks .

AllanMiller said...

There it is again - dot-dot-dot...

socol said...

Ok, Ok. I continue...
On the first and second paragraph of your response, I do not dispute the clock function , I want to highlight is the problem of assuming ( hypothesis ) that the differences between genomes of closely related species as indicators of evolutionary changes . Is absent evidence to say that only that percentage is responsible for the difference between one and the other species.
This hypothesis should be rejected when the results of applying this method of dating do not match those of the fossil record (assuming it is more accurate) .
I have clear that the method is not circular , but you claim that if their results offset by a "few" million years would make one calibration (this is a lifesaver to molecular method , when it should be re-evaluated the hypothesis that already mentioned) .
I believe that restricting the application of molecular method to modern species does not remove the problem I mentioned ; only minimizes the error becomes more evident when going backwards in the clade .
Regarding the third paragraph , I understand that the fossil with which the comparison is made must be a common ancestor before the node divergence ( hopefully it was the fossil node, but that does not happen ) .
However , this uncertainty increases the probability of failure in comparison dating. It is a problem that adds to those already mentioned.