WinAce

Taken from here:

Retroviral infections can occasionally infect a germ line cell. The resulting offspring will have bits and pieces of the virus stuck in every cell in its body. We've observed this very rare process in the lab, and the odds of getting two independent infections to leave the same bit of viral DNA at the same exact locus are astronomically unlikely.

Like pseudogenes, the viral fragments can piggyback on the success of an individual and become established in the species. The chances of any particular viral fragment, even if inserted identically in two seperate cases, becoming established in two seperate populations (a rare event in and of itself) makes this not just improbable, but more or less impossible without divine intervention

Any offshoot species will have the same unlikely and easily identifiable ERV, enabling us to construct accurate phylogenies from an independent line of evidence.



Exhibit A; human endogenous retrovirii insertions in identical chromosomal locations in various primates. Notice just how well the standard evolutionary phylogeny (humans and chimps closest, then orangutans and gorillas, then gibbons, then old world monkeys, then new world monkeys) is represented by this line of evidence.

ERVs have also been used to reconstruct the relationships between dogs, jackals, wolves and foxes; various breeds of domestic cat and wild cat; and even to establish the shared ancestry of cows and whales. (In the last case, two independent viral infections accounting for the evidence is impossible - whales and cows do not even share the same environment, much less are exposed to the same diseases!)

Needless to say, this offers numerous falsification avenues for evolution. Any ERV shared between organisms farther on the phylogenetic relationship than humans and apes must *also* be found in both. For example, ERVs found in New World Monkeys and chimps MUST be present in humans (aside from a few very rare cases where they've been deleted, but we can tell when a deletion has occured) or evolution is falsified.

An ERV in dogs and humans but not chimps would put the theory on its deathbed; so would a phylogeny reconstructed from these viral fragments if it differed significantly from the accepted phylogeny based on morphological, fossil, pseudogene, anatomical, and other evidence. This is the strongest support for evolution I've ever come across; a truly powerful and damning smoking gun.

Hypotheses proposed by creationists to account for this are inadequate: one entails independent insertion by the same virus affecting different species.

The fact that ERVs, when analyzed, yield evolutionary phylogenies consistent with the ones based on morphological, fossil and other evidence (as opposed to, say, viral infection patterns instead) rules out the independent insertion hypothesis. The fact that species that don't even share the same environment, much less diseases (like cows and whales) share unique ERVs falsifies it completely.

Another attempted explanation is that the ERVs found in our genome are actually designed elements originally placed there. Needless to say, this hypothesis is as ludicrous as assertions that dinosaur bones don't actually come from dinosaurs but were intentionally placed there. .

Science generally doesn't deal with Omphalos hypotheses for good reason, because when you deny reality in one area, you might as well say goblins created your mind on a supercomputer 5 minutes ago and this world isn't real.

People who assert this for ERVs have the burden of proof in demonstrating where shared ancestry ends and intentional design begins. Are the various breeds of domestic cat different kinds? (They share ERVs that have been used to reconstruct their phylogenies). Are the various felines related, or seperately created? (Lions, panthers, tigers and domestic cats share ERVs). In other words, where do the divinely faked ERVs end and the real ones begin?

Finally, ERV insertion is a well-documented event, leaving very specific and unlikely patterns. No other process except viral infection has been documented that can create them.

A customary red herring involved in any discussion of molecular evidence for evolution is cries of "but they have a FUNCTION!". While that may well be true, function (or lack of it) is generally not the criterion by which things are considered evidence for evolution.

In summary, the facts are that:

(A) viral infection generally results in a dead cell. Occassionally, fragments of the virus remain but the cell survives.
(B) infection of a germ line cell is even more rare.
(C) no two viral insertions are exactly alike.
(D) viral fragments insert at fairly random locations.
(E) that unlikely viral fragment-carrying germ line cell needs to be the particular egg or sperm cell that gets fertilized and survives to adulthood.
(F) It then needs to to piggyback on the success of an individual and get established in the population via sheer luck.

The combined odds of this happening in two independent "kinds" are probably orders of magnitude larger than those strawmen abiogenesis calculations creationists love to trot out. This leads to the obvious conclusion:

Independent origin of ERVs (and hence, independent origin of species, a.k.a. creationism) is scientifically falsified.

MattofVA

How does one know that a particular piece of DNA in chromosome is a virus and not part of the organism's inherent DNA?

Could one of these retroviruses go bad at some point and become harmful to the organism?

Other than that first possible question I would say that retroviruses are indeed a smoking gun!

judge

Retroviral infections can occasionally infect a germ line cell. The resulting offspring will have bits and pieces of the virus stuck in every cell in its body. We've observed this very rare process in the lab, and the odds of getting two independent infections to leave the same bit of viral DNA at the same exact locus are astronomically unlikely.


judge:
Can you explain exactly where we have seen this process "in the lab". Not in humans I presume?


The combined odds of this happening in two independent "kinds" are probably orders of magnitude larger than those strawmen abiogenesis calculations creationists love to trot out. This leads to the obvious conclusion:


judge:
Is it possible that you can be a bit more precise than "probably orders of magnitude larger"?

Thanks.

Doubting Didymus

Winace, I think this is absolutely superb. Great job. However, I'd also like to see one or two of the references you used, such as the retrovirus infections seen in the laboratory.

With respect to your probability calculations of course, I won't be asking. Because...:

The creationist probability arguments on abiogenesis are always bogus, firstly because we dont have exact data about what the first replicator was and how it arose yet, and neither do they, so calculating the probability of it is just silly. Secondly, they often completely ignore all the knowledge we do have about early replicators, and skip straight to calculating the probability of a whole eukaryotic cell, or even an entire organism, in a single step.

Given that all creationist abiogenesis probability calculations are plainly wrong, any attempt at comparing them with something else is equally worthless.

Judge: if you want to get an idea of the probability of independant insertion of these viral fragments, imagine first the simple probability of two occuring independantly in different species. First, the virus has to affect a germ cell, which is rare. Then the germ cell has to survive and gain viral fragments, rarer still. Then the germ cell has to be lucky enough to be included in the next generation. Then it has to overcome the odds of natural selection. This establishes a very low frequency for ANY retroviral inclusions in ANY species.

Then the two insertions have to be the exact same fragments in the two species.

Then the fragments have to be in exactly the same place in the two species genomes.

Then the individuals that have the inclusions have to overcome all other competition and pass their genes on, so that only their retrovirus infected genes can be seen in the whole population.

As you can see just by thinking on the matter, it's staggeringly unlikely that one viral fragment should occur in two unrelated species in exactly the same way.

Now, this has to happen not once, but once for every known retroviral fragment, and then in every species that has it. It has to do this each time EXACTLY the same as every other species. Your probability is being multiplied and multiplied and multiplied again.

WinAce

*Blink blink*

Retroviruses reproduce that way - they insert their RNA into host cells and hijack the cellular machinery. This is pretty basic stuff - do a search on "viral integration". You can introduce a virus into cell cultures in vitro and watch the ensuing consequences. This article explains it in very technical terms that went straight over my head. This one confirms what I've posted above:

"RNA retroviruses replicate by integrating a reverse transcript of their genes into the DNA of each infected host cell, which then makes new viral particles. If the host cell happens to belong to the germ line and survives infection, viral genes can be passed on to later generations, and this has happened many times."

In addition to Doubting Didymus' comments, the viral insertions need not just occur independently and identically at the same spots in different species, but do so in only the patterns common ancestry would impose. If, say, humans and chickens shared ERVs that apes didn't have, evolution would be in serious trouble. Since humans share a much closer environment with many domesticated animals other than the apes, the fact that this isn't found spells the death knell for independent insertion.

Mecha_Dude

That was exceptional WinAce, I only take note of one part where I'm sure creationists would try and nail you on:
The wording sorta blasts into your credibility even though it's scientifically sound. Maybe you could go into some detail on how the we can know when a deletion has occured. That might make it a bit more convincing to the reader.

WinAce

DNA deletion won't cut out the ERV directly. It's like parsing a string in a programming language... I'll try to show by example

Suppose we have two genomes (human and chimp).

Human:
ATTCGTACTTAGGGGCCCCGGGGTTATATGGT

Chimp
ATTCGTACTTAGGGGCCCCGGGGTTATATGGT

Very simplified. The red part is the viral insertion (not actual DNA bases that comprise it, just the basic idea).

Deletion mutations are random events. Now, let's see what happens when a deletion mutation affects the DNA of a chimp (three different ways):

Original:
ATTCGTACTTAGGGGCCCCGGGGTTATATGGT

Deletion 1:
ATTCGGGCCCCGGGGTTATATGGT

Deletion 2:
ATTCGTACTTAGCGGGGTTATATGGT

Deletion 3:
ATTCGTACTTTATATGGT

What happens? Well, the deletion doesn't just affect the ERV. It either leaves an identifiable bit of it or cuts it *and surrounding areas* out. In each case we can tell a deletion was at work.

But suppose the astronomically unlikely happens and the deletion specifically cuts out *only* the ERV, completely and without affecting its surroundings:

Plagiarized Errors and Molecular Genetics

"Other deletions that by chance do not remove any functional genes could eliminate some useless DNA including pseudogenes and retroposons; but an individual with such a deletion would have no particular selective advantage as a result of the deletion, so spread of DNA copies carrying the deletion into the population at large would be no more likely than the spread of any other inconsequential mutation."

The sum total of probabilities involved in not just removing the ERV, but doing so completely, missing any surrounding DNA, the mutation occuring in a germline cell that gets fertilized and survives to adulthood, *and* having the individual (with no selective advantage) overtake the population at large make it virtually impossible for an ERV deletion to mess up our studies.

Also, *bump* for Magus and other creationists.

ps418

WinAce,

There is an article in today's Science that discusses potential preferential intergration sites for retroviruses. As best as I can tell, this does not weaken the argument significantly, but does imply that the probability of retroviral integration is not precisely the same for every locus.

Wu et al, Transcription Start Regions in the Human Genome Are Favored Targets for MLV Integration. Science 300, pp. 1749-1751.




Trono, Virology: Picking the Right Spot. Science 300, 1670-1671.

Also:

Schroder et al, 2002. HIV-1 integration in the human genome favors active genes and local hotspots. Cell 110, 521-529.

EDIT: Adding one more article. This one should be available for free:

Elleder et al, 2002. Preferential integration of human immunodeficiency virus type 1 into genes, cytogenetic R bands and GC-rich DNA regions: insight from the human genome sequence. FEBS Letters 517, 285-286

EDIT2: Adding a great paper, also free, which appeared in the PNAS in 1999, dealing with the original argument:

Johnson and Coffin, 1999. Constructing primate phylogenies from ancient retrovirus sequences. PNAS 96, pp. 10254–10260.

Patrick

WinAce

Everyone's known about that for a while. Ashby Camp of TrueOrigins tries to use integration hotspots to ward off the obvious explanation, although he doesn't mention many details - and for good reason.

Assuming a hot spot exists that's 100,000 times more likely to be attacked by a virus than a regular section of DNA, over 100 times the hottest spot observed, that's still only a 1 in a 1000 chance of inserting there if you have 100 million base pairs to choose from. In addition, the abstracts of your articles appear to say that there are many preferred spots to choose from (as many as there are genes?), which, at most, cuts down the possible places where ERVs result from all DNA to a few places per distinct gene.

Relative numbers can sound convincing to the layman, but absolute ones blow the creationist explanation away.

Combined with the other factors, which hotspots don't even begin to address, this doesn't even make a tiny dent in the argument's force. But thanks for the info regardless =)