Stated Clearly: The DNA evidence for Common Descent in ERVs

[Mercer (John Mercer)]

@Rumraket

When a retrovirus enters a host cell and gets it’s own genome inserted into the host, it also needs it’s own retrovirus genome expressed.

I’m not following this. Retroviruses have their own, very strong, promoters in their long terminal repeats (LTRs). In addition to including cellular genes in their genomes, they integrate upstream of cellular oncogenes, the LTRs increasing their transcription and contributing to oncogenesis.

[Mercer (John Mercer)]

@Michael_Okoko

A nested hierarchy is like a family tree in some sense, because both show common descent, so I would use the concept of a family tree to demonstrate why nested hierarchies are good evidence for evolutionary.

I wouldn’t, because family trees are not nested hierarchies. It’s the nested aspect that creationists can’t bring themselves to acknowledge or understand.

[Rumraket (Mikkel R.)]

@Mercer

I’m not following this. Retroviruses have their own, very strong, promoters in their long terminal repeats (LTRs).

I said as much:

@Rumraket

When it comes to LTR retroviruses and retrotransposons, they even have canonical eukaryotic promoter regions, complete with enhancers, a TATA box, and initiators, in their LTRs.

[Michael_Okoko]

@Mercer

I wouldn’t, because family trees are not nested hierarchies

I didn’t say pedigrees are nested hierarchies, but like nested hierarchies they are most parsimoniously explained by common descent. The concepts of ancestors and descendants are important to both, and genomic data can be used to establish both.

@Mercer

It’s the nested aspect that creationists can’t bring themselves to acknowledge or understand.

Based on her comments, this appears to be the case, but I wasn’t trying to explain what a nested hierarchy is to her. I tried to illustrate why a nested hierarchy is best explained by common descent, based on her comment below:

@thoughtful

I understand that. I just don’t “get” why nested hierarchies are such good evidence for evolution.

If there is common descent, then evolution must have happened, so the presence of nested hierarchies (due to common descent) in biological sequence or morphological data constitutes clear evidence for evolution.

[John_Harshman]

@CrisprCAS9

If all of the sequences are identical, then it will form a hierarchy based on the rest of the sequence, since the only necessarily constant region in my hypothetical was the spacer.

But you were constructing your trees from the spacer, weren’t you? Doesn’t matter. Let’s just suppose that the rest of the protein is also identical among species. Why not?

@CrisprCAS9

It will resolve to a nested hierarchy when you run the data through the algorithms, which was my point.

Bad point, if so. I say again: just the existence of a single “best” tree (which will only happen with likelihood, not parsimony) is not a nested hierarchy in any but the most trivial and arbitrary sense, just like that “phylogeny” of cars we keep running into. It’s significant hierarchy in the data themselves, detectable through statistical tests, that counts.

@CrisprCAS9

The creationists are claiming that you can’t infer a ‘real’ hierarchy because the algorithms will resolve any data set into an ‘apparent’ nested hierarchy.

In this they are correct. Nested hierarchy in the data can’t be inferred from the simple existence of a tree. Either inner consistency, as shown by significance in a test for signal, or mutual consistency of independent samples, or both, is necessary.

[thoughtful (Valerie (non-scientist))]

@Faizal_Ali

And in the video that is the subject of the thread, they calculated the odds against the ERV insertions coinciding with what would be expected under common ancestry just by chance to be astronomical - for just two species . The odds of this being observed across every single species that we have looked at, which is the case, are even lower by many orders of magnitude.

I may be opening a can of worms by saying this, but I find it ironical that abiogenesis (according to Tour) and even evolutionary history as a whole (some threads on the forum) are shown to have similar astronomical odds, but that doesn’t deter many scientists from thinking they occurred without design or a creator.

@Faizal_Ali

Creation does not make such a prediction.

But creation makes a prediction of function. And we are just beginning to find how they are functional. So then the identical placement of insertions explanation has no merit, if these were created sequences, except in the nested hierarchy argument perhaps that they were describing below.

@Rumraket Thanks for the explanation and links - I just need to learn this biology better to be clear on what you’re saying and probably had some misunderstanding what that article meant because the explanations don’t seem to match up to me. I’ll have to bookmark this stuff.

@CrisprCAS9 and @John_Harshman Thanks for explanations as well. I will bookmark, but I’m struggling to understand without a visual as to the what the hierarchies look like in each example that Crispr gave, especially the ones that we do actually see in sequences that for scientists point to common ancestry.

[Rumraket (Mikkel R.)]

@thoughtful

@Rumraket Thanks for the explanation and links - I just need to learn this biology better to be clear on what you’re saying and probably had some misunderstanding what that article meant because the explanations don’t seem to match up to me. I’ll have to bookmark this stuff.

Yes it’s not the same as I take the authors to answer more specifically about why a retrovirus that infects immune cells specifically (like HIV does) would induce regulation of immune-specific host genes, and the answer I give applies more broadly to the relationship between DNA and DNA-binding proteins in retroviruses and retrotransposons in general (why is it that host DNA binding proteins can bind and regulate both integrated virus and host DNA?).

They give an evolutionary (selection) reason why retroviruses that infect immune cells would have this effect on host immune system genes, while I give a general biochemical reason why such associations between DNA and protein is both unavoidable, and a selective reason why recruiting host transcription factors is in the “interest” of retroviruses more broadly.

[CrisprCAS9]

@John_Harshman

In this they are correct. Nested hierarchy in the data can’t be inferred from the simple existence of a tree. Either inner consistency, as shown by significance in a test for signal, or mutual consistency of independent samples, or both, is necessary.

You realize that was my entire point, right?

[Chad_the_Layman (Chad Stewart)]

@thoughtful

I may be opening a can of worms by saying this, but I find it ironical that abiogenesis (according to Tour) and even evolutionary history as a whole (some threads on the forum) are shown to have similar astronomical odds, but that doesn’t deter many scientists from thinking they occurred without design or a creator.

I don’t think this question opens a can of worms. And I think the answer is important. I’m going to take a shot and see if I get corrected, err educated.

If you stand beside LUCA and place a bet on what she will become in 4 billion years, your odds of winning that bet are astronomical.

If you stand beside me and place a bet on whether LUCA is my ancester, you might make a little bit of money. That is probably what happened.

Odds are different depending on where (or when) you stand.

The odds discussed in the video are the odds that ERVs found in the same exact place in the genomes of different species occured randomly instead of via common descent.

And I can’t think of a way to relate that to my betting on LUCA analogies, LOL.

Well, maybe this is a helpful fail.

[John_Harshman]

Here, try this. Look at figure 2. Don’t worry two much about anything except the solid bars running through the branches, each of which is a change in sequence shared exclusively by the descendants of that branch. There are a very few unfilled bars, representing changes that don’t fit the tree. That’s what nested hierarchy looks like.

Any questions?

[John_Harshman]

@CrisprCAS9

You realize that was my entire point, right?

No, because that isn’t what you said. I will accept that it may be what you meant to say.

[Mercer (John Mercer)]

@John_Harshman

Here, try this .

Why MYC? I see the listed reasons, but at least in mammals it is very G-C rich. Does that somehow make it good for phylogenetics?

[thoughtful (Valerie (non-scientist))]

@John_Harshman

Here, try this . Look at figure 2. Don’t worry two much about anything except the solid bars running through the branches, each of which is a change in sequence shared exclusively by the descendants of that branch. There are a very few unfilled bars, representing changes that don’t fit the tree. That’s what nested hierarchy looks like.

Cool. Thanks for linking me to a paper you co-authored. Easy enough to see there.

@John_Harshman

Any questions?

Do you know of a paper of a figure like that with organisms that creationists would not consider created kinds? Even apes and humans. I’m not sure if I’ve come across one yet. But I’d be interested if there were other examples as well.

[Faizal_Ali]

@thoughtful

I may be opening a can of worms by saying this, but I find it ironical that abiogenesis (according to Tour) and even evolutionary history as a whole (some threads on the forum) are shown to have similar astronomical odds, but that doesn’t deter many scientists from thinking they occurred without design or a creator.

The difference is that, in the case of ERV’s, we have an explanation that makes the observed pattern, not only probable, but inevitable. So it would be perversely irrational to, instead, prefer an explanation the odds against which runs to hundreds of digits long.

@thoughtful

But creation makes a prediction of function.

And that prediction has been has been completely, unambiguously falsified. That you are unable to understand or accept this fact is not an argument against it, sorry.

@Chad_the_Layman

The odds discussed in the video are the odds that ERVs found in the same exact place in the genomes of different species occurred randomly instead of via common descent.

And, to avoid further confusion, since these sequences are nonfunctional “design” is not an explanation at all for the pattern of insertions. However, even if all “kinds” were separately created by a God (including the ERV “kind”), insertions would still have occurred, but not in a nested hierarchy. They would be random, so in that respect ruling out random insertion is still a test of design/creation.

[Rumraket (Mikkel R.)]

@thoughtful

I may be opening a can of worms by saying this, but I find it ironical that abiogenesis (according to Tour) and even evolutionary history as a whole (some threads on the forum) are shown to have similar astronomical odds, but that doesn’t deter many scientists from thinking they occurred without design or a creator.

Right, but many of those odds calculations are disputed. It’s not that scientists agree “yeah that’s the real number” but then just ignore it and think life originated and evolved anyway. They don’t think the numbers are accurate.

I’m not aware that anyone disputes that those are the odds of that many ERV insertions in the same locations for ERVs that randomly insert themselves, though I have heard it claimed that ERV insertions aren’t random.

[Rumraket (Mikkel R.)]

@thoughtful

Do you know of a paper of a figure like that with organisms that creationists would not consider created kinds? Even apes and humans. I’m not sure if I’ve come across one yet. But I’d be interested if there were other examples as well.

What you’re seeing on that figure can be done with any imaginable gene sequence shared among the species used to depict the phylogeny. Most phylogenetic trees published usually just show the branching order, branch lengths, and node support values. It is a trivial (but rather tedious) exercise to depict the changes that occur on branches because for large phylogenies and long sequences there’s so many of them.

[John_Harshman]

@Mercer

I see the listed reasons, but at least in mammals it is very G-C rich. Does that somehow make it good for phylogenetics?

The part we sequenced isn’t GC rich, at least in crocs and birds, and in fact the GC rich part is difficult to sequence. We made a couple of attempts at it but couldn’t get good results.

@thoughtful

Do you know of a paper of a figure like that with organisms that creationists would not consider created kinds?

That’s a problem, because nobody has any clear idea of what created kinds are. Do you think crocodylians are all a single kind? If so, I can easily find creationists who don’t. And what does that matter, anyway? You asked only for a demonstration of nested hierarchy. Now that you understand what one is, you can apply it elsewhere.

@thoughtful

But I’d be interested if there were other examples as well.

Do you mean examples of graphics like Fig. 2? There are others, but again, I can’t tell what you think are single kinds and what are multiple kinds, so I can’t offer you anything. (That should be instructive, by the way: if there really were separate kinds, they ought to be easy to see. That they aren’t is evidence against the whole idea.)

[Faizal_Ali]

@Rumraket

I’m not aware that anyone disputes that those are the odds of that many ERV insertions in the same locations for ERVs that randomly insert themselves, though I have heard it claimed that ERV insertions aren’t random.

It seems to me the video accommodates the possibility that insertion is not completely random by saying a particular virus could have inserted into any of 10 million sites, which would be only 0.3% of the genome. I hope I understood that correctly.

[John_Harshman]

Yes, insertion is not completely random, if by that is meant insertion is equally likely between any two adjacent sites. The neighborhood of frequently expressed genes is over-represented. On the other hand, by the time we see insertions in the genome the deleterious insertions have been weeded out by selection, so that’s another constraint. There are a few known cases of homoplasious insertions at exactly the same site, but few enough to show that it’s a very rare event.