XMRV Evilution: Its more than just change over time

Well done explanation.

Look by divine Law XMRV is precisely 6015 Years old and any one who blasphemes against the word of God Mikovits will go to hell.

I seem to be unable to grab the Provisional PDF from that link, whenever I download it it ends up as a broken 50kb file. I had a question about the methods used in the phylogenetic reconstruction. Specifically how they did their alignments (which program) and what program and evolutionary model they used for their phylogenetic reconstructions?

But aren't you looking at the wrong virus here?

If you look at HTLV it does the same thing.

"The evolutionary rate of the human T-cell lymphotropic virus type-1 (HTLV-1) is considered to be very low, in strong contrast to the related human retrovirus HIV."
http://mbe.oxfordjournals.org/content/21/3/603.full.pdf

" In agreement with experimental findings, it is assumed that cell activation is required for successful replication in T helper cells and that HTLV-I induces a significant degree of bystander activation. It is found that the rate of evolution of HTLV-I is limited by the restricted availability of activated uninfected T cells, both at high and low proviral loads. This limits the within-host sequence diversity of HTLV-I and may therefore account for the slow rate of evolution of the virus in the population."
http://www.ncbi.nlm.nih.gov/pubmed/10824088

"DNA analyses of HTLV-I sequences demonstrated that (i) little or no genetic variation occurred in vivo in the same individual or in different hosts from the same region carrying the same virus, regardless of their clinical statuses"
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC289023/pdf/jvirol00166-0474…

I'm looking at the pdf now Dan. The link is all jacked up and has a / instead of a . before the pdf. There's no methods section which means I'll have to read it :(

They don't say. They cite another paper for the trees and data (Paprotka, T., K.A. Delviks-Frankenberry, O. Cingöz, A. Martinez, H.-J. Kung, C.G. Tepper, W.-S. Hu, M.J. Fivash, J.M. Coffin and V.K. Pathak, Recombinant
Origin of the Retrovirus XMRV. Science, 2011. 333(6038): p. 97-101.)

They do say this, hehe: "Rather than evolving as would be expected of a true infecting virus, these sequences therefore seem to be simply moving up and down the MLV phylogenetic tree. The conclusion that they represent MLV sequences amplified from trace mouse DNA contamination is inescapable."

Coffin is avoiding the use of scientific terms here and is is really making political statement.

jel-- This paper is not on mutation rates. Nothing to do with mutation rates. It has to do with the actual mutations. Yes, things change over time, but a population of elephants doesnt evolve into mammoths, and then evolve into chickens. You dont evolve into something already in existence. There is no logical reason why an exogenous, actively replicating retrovirus be identical to one ERV one day, and another ERV the next. It makes more sense that a lab is seeing contamination, and that they are detecting MPMV1 one day, and PMV6 the next.

@Jel

It really doesn't matter if there is much (like HIV) or little (like HTLV-1) variation: phylogenetic analysis shows you which sequences are ancestral to other sequences and which sequences are not. Lo's *old* sequences are not ancestral to the *new* sequences FOUND IN THE SAME PATIENT according to phylogenetic analysis.

This pretty much proves Lo's findings are contamination. The authors state "the conclusion that [the Lo sequences] represent MLV sequences amplified from trace mouse DNA contamination is inescapable".

Again a good, insightful post and I must say that again, you called it (right away), ERV.

One thing that bugs me is the simplicity of this analysis. We all recall that the reviewers asked Alter/Lo to show integration sites, but why didn't they demand a phylogentic tree? It could have solidified their case if they we're right and prevented to provide false hope if they were wrong (as it now appears they clearly were). Why would it take close to a year for someone to perform this (seemingly) extremely straightforward analysis?

jel:

Remember that the tree with the ERVs added the MLV-Like "patient" samples are branching within endogenous mouse retroviral clades. That isn't the result of slow divergence and relatively low rates of sequence evolution. Unless what you really expect to see CFS patients having been infected with related but distinct viruses that have undergone parallel switches to a human host.

Granted I don't know a lot about the taxonomy of MLV-Like viruses, I know some can infect multiple vertebrates but these trees don't look very good to me from a "MLV-Like viruses cause CFS" point of view. The similarity of the "control" tree where contamination is definitely the culprit is pretty telling.

ERV: Have you commented on the Hanson work that supposedly showed infectious particles in patient blood? It looks like they saw similar results as Lo et al but did more controls. Unfortunately when I poke around it seems like all I can find is a conference abstract in Retrovirology and not a full paper yet. Just wondering if maybe you had heard more details.

@ Dan

I can give you some more details.

Hanson first presented her results at the September 2010 Workshop on XMRV.

http://regist2.virology-education.com/abstractbook/2010_8.pdf

Check abstract O_11.

She reported positives in 8/10 "severe CFS cases" 3/10 "recovered CFS patients" and 1/10 "healthy, non-household controls". Note that while the abstract book mentions XMRV, at the Workshop Hanson explained that this was a mistake on their part and that they had found that these sequences were actually more like Lo's sequences.

At the Blood Products Advisory Committee meeting in December of 2010, Hanson presented again:

http://www.fda.gov/AdvisoryCommittees/CommitteesMeetingMaterials/BloodV…

See page 197 and further.

Curiously, she added 10 controls after unblinding for the September Workshop. She now finds 7/10 "severe patients" (which seems odd because it earlier were 8/10), 7/10 "recovered patients" and 4/20 controls to be positive.

I am not a virologist, but the following quote from Hanson seems rather worrying, especially in light of the recent Shin et al. study:

"We have three different brands of PCR machines in our lab. We have seen differences in which PCR machine is used, in our own lab, as to what results we get as far as detecting things or not detecting things."

Finally, as it doesn't seem that Hanson collected and handled all samples in the same manner (the addition of 10 controls alone kinda proves that), it seems prudent to treat these results (even if published) with caution. Like Cingöz and Coffin explain quite clearly in their paper:

"A second, related issue regards the provision of appropriate controls. Given the apparent sporadic nature of this sort of contamination, it is absolutely essential that ontrols and patient samples be exactly matched, not only for personal characteristics, such as age, gender, geography, etc., but also in the reagents and materials (tubes, needles, etc.) used to obtain and process the assay samples. Unfortunately, such caution is often not the case, particularly in retrospective studies like Lo et al. [14]. Clinical samples are often collected as at least two separate groups, the simplest example being patient vs. control samples. This lack of caution can result in detection of a contaminant in one set of samples and not the other, resulting in false association of virus with human disease."

JohnV: Thanks for the tip, managed to get the PDF now, not much to it really. Looks like they used ClustalW for the alignments and MrBayes for the trees. Wouldn't be my prefered method. ClustalW is probably fine for the alignments, I don't work with viral data so I don't know how easy/difficult these sequences are to align, but I would have preferred an ML Tree myself over a Bayesian tree. But I suspect the result in this case would be basically the same.

RPM: I don't think there are samples here from the same patients at two different time points. I believe the mid-90's and 2010 CFS patient samples are two distinct cohorts. The resulting Phylogeny is still fucked up though from the CFS = MRV infection viewpoint. Contamination really does look like a culprit here. But the difference in contamination between patient and healthy controls is a little troubling. Could just be random though.

@Dan

No, you are wrong. Lo and Alter revisted some of the same patients (the ones they could track down) for their PNAS paper. Thus, these sequences do include samples from the same patients at two different time points.

Hey - you are making some rather loose assumptions about the rate of evolutionary changes and yourself are working with rather pronounced prejudices.

why don't we all just wait for Lipkin to finish his research and see what he says about the whole XMRV / MLV-like virus questions.

or do you think he has an axe to grind or money to make too and that you know more than he does about virus identification and such.

@RPM: Thanks for the Rundown of the Hanson stuff. I absolutely agree, if you notice differences between PCR machines (something I actually don't find terribly surprising, the thermal cycling performance of different machines can be quite different and if we are talking degenerate primers here that will make a big difference in amplification) you need to be treating all samples identically. Your detection protocols have to be identical. Reagants, machines, handling, etc start to finish.

And my bad on the second part, I only skimmed the Lo paper and assumed that it was a new cohort based on their description in the abstract.

... loose assumptions about the rate of evolutionary changes...
This has nothing to do with rates of evolutionary changes.

Chickens dont evolve into ladybugs.

@Edugreat Assumptions about the evolutionary rate don't affect the interpretation of those phylogenetic trees unless you believe that the rate makes the reconstruction invalid (model misspecification). If rates were low enough that the branches were all unreasonably short then we may see artifacts due to long-branch attraction where sequences with higher relative rates to the rest of the dataset go together simply because they are more divergent overall and not because they are actually more related.

From an eyeball of those trees, that doesn't really seem to be the case. Either way those sorts of questions are actually relatively easy to test.

Evolution: a theory that the various types of animals and plants (OR VIRUS) have their origin in other preexisting types and that the distinguishable differences are due to modifications in successive generations (TIME);

To clarify my statement about rate of change, In classical evolution you would expect the differences between P1s, F1s and F2s to increase as TIME goes on - Correct? so you are making the assumption that because YOU don't see the changes YOU assume should happen at the rate YOU think is should then the results are invalid. You assumptions must be based on some idea of what it should do compared to what? HIV-1?
What if these viruses are more stable then YOUR preconceived notion?
That is what I mean when I say that "you are making some rather loose assumptions about the rate of evolutionary changes and yourself are working with rather pronounced prejudices!"

Wait for Lipkin!

... so you are making the assumption that because YOU don't see the changes YOU assume should happen at the rate YOU think is should then the results are invalid. You assumptions must be based on some idea of what it should do compared to what? HIV-1?

This has nothing to do with mutation rates. It has to do with the actual mutations. There are not only one dozen functional gag sequences, thus viruses just cycle between the 12. There is no reason why a modern virus should have the exact same sequence as an ERV generated thousands to millions of years ago, then change to another sequence that just happens to be identical to an ERV generated thousands to millions of years ago.

Wonder why John Coffin would say this about a contaminent to the Globe (how could a contaminent cause a small infection):

Reporters Question: Could this XMRV, which you say developed in laboratory mice, have escaped and be causing different problems in people, if not chronic fatigue?

Coffin's Answer: We donât think it could survive well. It conceivably initiates a small infection [but] we think itâs very unlikely to initiate a major infection.

@ RRM (not R.R. Mendoza)8

Sorry but that confirmation raised a smile. The 'debate' over your 'true identity' was laughable.

Good on you :)

Why would the researcher(s) do that? How does it benefit them, and why is it so important for them to illustrate MLV infects humans that they would go out of their way to contaminate data? Are they getting paid by interested parties to establish such a link, or...?

I just don't understand how such dishonesty stands to benefit those who did it.

0verlord-- CRAP! No! No one was purposefully misrepresenting data!!

What happened was, is that Coffins group wanted to know how much mouse DNA would you need to actually be there for contamination to be a real issue (the issue of contamination has been a big one in XMRV/MLV-whatever research). *They* purposefully 'contaminated' reactions to find out that as little as 1/100th of one mouse cell could screw you over, thus labs doing XMRV/MLV-whatever research need to be EXTREMELY careful.

Independently, a lab published sequences they 'found' in CFS patients and blood donors.

When Coffins lab compared their data to the other labs data, they were like 'It looked like that other labs data was contamination before. Upon further analysis, it definitely was, and considering you only have to have 1/100th of a mouse cell around to get these kinds of results, PEOPLE NEED TO BE REALLY FRIGGEN CAREFUL!!!'

No one was purposefully misrepresenting data.

But dont feel too badly for the lab being outed as being 'careless' in this paper, though. This could have been avoided, but they used the PNAS publishing system to avoid standard peer review. They deserve to be publicly chastised like this, and ultimately they should retract their paper.

erv, Wow, talk about my missing the point! Silly me. Now I understand, though. Thanks for the explanation, that makes a lot more sense.

I am familiar with evolution, but I had not yet heard about evilution.

Is this some new process that is used by WPI?

WPI is not practising EVIL lution, but I believe John Coffin is. Interesting Freudian slip.

Ernie-- There are numerous sources of contamination with these XMRV/MLV-whatever papers. This paper is addressing mouse genomic DNA contamination. Coffin in that article is addressing cell line contamination (a very common HIV-1 indicator cell line is contaminated with MLV).

@Ernie

That Coffin answer really shows the level of understanding from the forums again.

Coffin is reffering to the fact that, if you would somehow "get" XMRV in human blood, even then you would not expect any problems within that human. He's aluding to the fact that although XMRV spreads in some human cell lines very well, human APOBEC3 inhibits XMRV infectivity.

Thus, XMRV can potentially 'initiate a small infection but it is very unlikely to initiate a major infection.' And therefore there is also nothing wrong about saying this about a contaminant. In fact, it's exactly what you would expect from a contaminant that originated in a human cell line: it can survive very well in that human cell line, but not that well in a new environment: "true" human material.

Edugreat, this paper has nothing to do with the rate of evolution over time, whether it be great like HIV or small like HTLV. What this paper deals with is that, according to the phylogenetic tree devised by Coffin, the mutations reported by Alter and Lo are not consistant with viral evolution at all, big or small, but rather are consistant with sample contamination. For the mutations to be consistant with viral evolution, the 2010 sequences should have to come from, or at least be related to in some way, the earlier mid-nineties sequences, ie the later sequences should branch off the earlier sequences, and not simply 'jumping up and down the tree' as stated by Coffin.

As for the purposeful contamination tree, that's just brilliant and Mary Kearney did that. Dr. Kearney has been involved with lots of basic covering the bases work with XMRV and there's something about her work that I really like. It's just so step by step that it's neat to read.

What about infection with multiple MLV's, as is observed in other other mammals?

@Lisa

Remember, from the original samples, Alter/Lo sequenced 21 PCR products and 18 of them (!) were identical (CFS Type 1 in the tree). It doesn't make any sense to assume that Alter/Lo first found all these identical MLV's (and none of the "other" MLV's) and later on, in the same patients using the same primers, only the "other" MLV's and no "CFS-Type 1 MLV" or its descendents.

Alter/Lo claimed that this added experiment was supportive of their finding (because the sequences "evolved" over time) and that claim in their paper has been shown to be false.

The link to the PDF at the journal's webpage is still borked, this one seems to work:
http://www.hindawi.com/journals/av/aip/940210.pdf

Thanks for the topic discussion ERV as a layperson I do find some (most :)) of the associated papers confusing at times.

I am learning though and that is most welcome :)

Thanks for the topic discussion ERV.

As a layperson and patient I do struggle with understanding some (most ;)) research papers and interpretations.

But I am learning and that is most welcome :)

Studies have show that HTLV does not change over generations. So what are you talking about? This is not evidence against at all.

@Terry

What are YOU talking about?

The Lo study showed evidence that the MLV-like viruses in CFS patients, DID change over generations. Only, now that other scientists have done the experiments that Lo/Alter should have done themselves in the first place, we know that the "evolved" sequences are in fact NOT related to the supposedly ancestral sequences.

The HTLV hypothesis is about the lack of sequence diversity. This blog post is not about sequence diversity at all, but about the phylogenetic relationship between different sequences that were found in the same patients over a period of time.

Thus, the (ad hoc) HTLV hypothesis does not "save" the Lo/Alter paper or "refute" the Coffin/Cingöz findings.

I should not be commenting on this article as it is gooblie gob to me. I have been ill for some time & have a multitude of diagnoses, thus little faith in the medical community any longer.

I want to ask a really stupid question. Do herpesviruses have a similar tree? What would it mean if when tested for several I showed acute EBV but the next time I was tested I showed acute CMV & the next time it was shingles & so on?

@#34 Terry, "Studies have show that HTLV does not change over generations. So what are you talking about? This is not evidence against at all."

The fact that HTLV does not change or changes very little over time has nothing to do with this paper or the original paper by Alter/Lo itself.

While waiting for their paper to be published as a result of it being held up in an effort to try and reconcile their findings with the CDC's negative findings, Alter and Lo located and re-tested 8 patients from the original cohort of mid-nineties samples. They found 7 of 8 to be positive again, except with different sequences than were reported originally. Alter and Lo took this sequence variation to mean that the virus supposedly infecting their patients had mutated over time.

What Coffin and Cingoz did in the above paper was to re-examine these sequences, except that according to their phylogenetic tree as shown above, the sequences amplified from the 2010 samples could not have come from the sequences amplified from the mid-nineties samples taken from the exact same individuals, indicating that instead of finding evidence of viral evolution as previously claimed, Alter and Lo had instead simply amplified new sources of contamination.

Abbie wrote: "You are different from your parents."

I have stopped several creationists dead in their tracks by asking them if they are identical to both their father and their mother (think about it - there's a problem). If they are not, therefore evolution - change over time.

RRM - what ancestral sequences?

HTLV does not have to mutate either.

http://www.ncbi.nlm.nih.gov/pubmed/8372432

Virology. 1993 Oct;196(2):506-13.
Interfamilial and intrafamilial genomic diversity and molecular phylogeny of human T-cell lymphotropic virus type I from Papua New Guinea and the Solomon Islands.
Nerurkar VR, Song KJ, Saitou N, Melland RR, Yanagihara R.
SourceLaboratory of Central Nervous System Studies, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland 20892.

Abstract
To determine the interstrain genomic diversity and molecular phylogeny of the recently identified variants of human T-cell lymphotropic virus type I (HTLV-I) in Melanesia, we enzymatically amplified, then directly sequenced representative regions of the gag, pol, and env genes of HTLV-I strains from 10 members of four families, including one family from Papua New Guinea and three families from the Solomon Islands. When aligned and compared to a Japanese strain of HTLV-I (ATK), the Melanesian HTLV-I strains differed by 7.6 to 8.7% in the gag, 7.1 to 9.3% in the pol, and 7.3 to 8.2% in the env gene regions. Based on 931 nucleotides, the overall sequence divergence of the 10 Melanesian HTLV-I strains from HTLV-I ATK was 7.3 to 8.1% (68 to 75 base substitutions). The intrafamilial genetic heterogeneity among these virus strains was nil to 0.2%, while the interfamilial sequence variation between HTLV-I strains from the Solomon Islands and those from Papua New Guinea was 3.4 to 4.2%, and the genetic heterogeneity among virus strains from the three Solomon Islands families was 0.2 to 0.9%. Using the maximum parsimony and neighbor-joining methods, phylogenetic analysis indicated that the HTLV-I strains from Papua New Guinea and the Solomon Islands formed a monophyletic group and that the Melanesian and cosmopolitan strains of HTLV-I have evolved along two major geographically dependent lineages.
PMID: 8372432 [PubMed - indexed for MEDLINE]

http://www.ncbi.nlm.nih.gov/pubmed/7826698

AIDS Res Hum Retroviruses. 1994 Sep;10(9):1135-42.
Familial transmission and minimal sequence variability of human T-lymphotropic virus type I (HTLV-I) in Zaire.
Liu HF, Vandamme AM, Kazadi K, Carton H, Desmyter J, Goubau P.
SourceRega Institute for Medical Research and University Hospitals, Katholieke Universiteit Leuven, Belgium.
Abstract
Our group previously reported a strong familial clustering of HTLV-I-associated myelopathy/tropical spastic paraparesis (HAM/TSP) in Zaire, suggesting a familial transmission of the virus together with the presence of cofactors. In the present study among 84 relatives of 16 HTLV-I-positive or HAM/TSP index cases, we found that all 15 seropositive children had a seropositive mother and that all 15 children with a seropositive father but a seronegative mother were seronegative. Lymphocytes of 17 relatives from 2 families with a familial HTLV-I-associated neuropathy were tested in 2 polymerase chain reaction (PCR) assays amplifying pol and tax/rex gene fragments. The 10 seropositive individuals were PCR positive for HTLV-I and the 7 seronegatives were negative in both PCR assays. The PCR results showed no evidence for a long lag period between infection with HTLV-I and seroconversion. The HTLV-I long terminal repeat (LTR) of these 10 individuals, related in the first to the fourth degree, was amplified and sequenced. Identical sequences were found within the families except for one woman infected with two variants, one being the familial strain and the other a mutated one with a single nucleotide substitution in the 755 sequenced nucleotides of the LTR region. The family strain and the mutant were both present in two samples taken 1 year apart. Together, the HTLV-I serology, PCR, and sequencing results point toward mother-to-child transmission as the main mode of HTLV-I infection in this population. Comparison of the LTR sequences of the two families with other HTLV-I strains from different geographical regions shows that the Zairean HTLV-I strains form a separate cluster.(ABSTRACT TRUNCATED AT 250 WORDS)

ERV how does one prevent/detect mouse contamination? Can it ever be ruled out completely and will every piece of research be 'suspect'?

Sorry - these thoughts have been 'bugging' me :)

@39:
extremely tiny, like very, very tiny, mousetraps. They're powered with microphants, for truth!
See this scholarly paper on it:
http://elephanticity.250x.com/thetruth.html

@Jack

Although ERV will no doubt provide a better answer, I think the answer is that with MLV's (unlike XMRV itself), showing actual human integration sites would go a long way toward ruling out contamination.

Apart from that, you really just have to treat ALL samples in the EXACT same manner from start to finish (which none of the positive studies have done thus far): any sort of (mouse) contamination should then show up equally in patients and controls. As the authors of the paper that's the subject of this blog post state:

"Given the apparent sporadic nature of this sort of contamination, it is absolutely essential that controls and patient samples be exactly matched, not only for personal characteristics, such as age, gender, geography, etc., but also in the reagents and materials (tubes, needles, etc.) used to obtain and process the assay samples."

Note that XMRV contamination from (for examples) 22Rv1 is entirely undetectable by screening for mouse DNA (and that showing human integration sites would not be that convincing as 22Rv1 is after all a human cell line). Therefore, collecting and handling everything in exactly the same manner is really the only way to go with any "is XMRV in humans?" research.

Jack, I'm just a patient who has been following the research and not a scientist (and therefore couldn't tell you much more than what's stated below) but I think along with different contamination tests, including RRM's response above about everything needing to be done in the exact same manner, same time, same place, same people, etc., that's also where confirmatory studies come in. Once something is detected, there are lots of 'and then' studies to do, for example-

1. Antibody studies- Isolate the virus and inject animals with it and see if/what antibodies they make. Take these antibodies and run them against your human samples and see if they react. Also you can kill and dissect the animals to see what kind of tissue tropism occurs as well as see if the animals get sick from being injected.

2. Sequencing the PCR products- Resulting PCR products should have their entire genomes fully sequenced. If the fully sequenced viruses all relate to another a logical manner instead of being a) basically the same virus, as occurs with XMRV, or b) bopped around all over the place and not even being the same virus, as occurs with the MLV sequences evidenced above, then that also works in one's favor.

3. Infect human cell lines with the virus and see what types of cells it likes to infect- (T cells, NK cells, B cells, etc). See if the virus establishes a productive infection and then test patients and see if the same types of cells are missing or otherwise farked up in some way.

4. Determine what kind of in vivo restriction factors are at play (APOBEC, etc.) and see if the patients in the studies are deficient in those factors or if those factors instead efficiently restrict infection. In XMRV several restriction factors have been reported to efficiently control XMRV infection, indicating that natural infections to occur would be somewhat difficult.

5. If it's a potentially zoonotic virus, study different animals and see where the virus came from.

The thing is, basically all of the above studies have been done in regards to XMRV and all of them have pretty much been dead ends. It's not just about 'politics' and other labs not calling Judy Mikovits to personally do their tests for them as certain parties are known to assert, it's the totality of the research base. Basically on one side there's the WPI's paper for XMRV and on the other side there's the rest of the research literature against XMRV.

I think the WPI still has some sort of Hail Mary in regards to their antibody findings (which might or might not even end up relating to XMRV for all I know), with Alter also commenting in some interview or another that most of his patient samples were also positive for antibodies while most of the controls were negative (not sure what kind of antibodies he used though), but I guess we'll just have to see how that plays out.

@Jel

Alter/Lo found a different strain of MLV in 7 out of 8 patients they revisited several years later. They took this as evidence for their findings, as the MLV's had supposedly evolved. Thus, Alter/Lo asserted that the earlier strains were ancestral to the later strains, not me (just check the paper, it's right there).

This paper shows that Lo/Alter's assertion is false, as the earlier found strains are not ancestral to the later found strains. The earlier found strains could realistically not have evolved into the later found strains.

It's not a matter of 'not having to mutate'. as it simply isn't a matter of sequence diversity but of phylogenetics that undermines Alter/Lo's findings.

Thanks RRM and Jon Oh and Justicar too for taking the time to reply so well. You certainly make me think - always a good result I believe :)