• Welcome to Phoenix Rising!

    Created in 2008, Phoenix Rising is the largest and oldest forum dedicated to furthering the understanding of and finding treatments for complex chronic illnesses such as chronic fatigue syndrome (ME/CFS), fibromyalgia (FM), long COVID, postural orthostatic tachycardia syndrome (POTS), mast cell activation syndrome (MCAS), and allied diseases.

    To become a member, simply click the Register button at the top right.

Cleaning up after XMRV

Ecoclimber

Senior Member
Messages
1,011
http://www.virology.ws/2012/05/18/cleaning-up-after-xmrv/
The retrovirus XMRV does not cause prostate cancer or chronic fatigue syndrome – that hypothesis was disproved by the finding that the virus was produced in the laboratory in the 1990s by passage of a prostate tumor in nude mice. A trio of new papers on the virus attempt to address questions about the serological detection of XMRV in prostate cancer, and further emphasize that XMRV is not a human pathogen....
 
Messages
5,238
Location
Sofa, UK
http://www.virology.ws/2012/05/18/cleaning-up-after-xmrv/
The retrovirus XMRV does not cause prostate cancer or chronic fatigue syndrome – that hypothesis was disproved by the finding that the virus was produced in the laboratory in the 1990s by passage of a prostate tumor in nude mice.
Still don't get the logic of this argument. Reads a bit like "It was made in the laboratory, so it can't be harmful". How can we confidently assert what XMRV does not do at this stage?
 

FancyMyBlood

Senior Member
Messages
189
Still don't get the logic of this argument. Reads a bit like "It was made in the laboratory, so it can't be harmful". How can we confidently assert what XMRV does not do at this stage?

If you click the link in 'produced in the laboratory'' it states "Because XMRV arose between 1993-1996, this scenario could not explain cases of prostate cancer and chronic fatigue syndrome that arose prior to that date.''

It doesn't say XMRV can't be harmful, but obviously it's impossible to be a cause of ME/CFS or prostate cancer because of these findings and the fact that XMRV is not found in the human population.
 
Messages
13,774
I guess that you can argue that if rates of PC and CFS have not dramatically increased since 1993-1996, then it's unlikely that XMRV causes them, or a significant number of them.

To me, particularly with CFS where prevalence, cause, etc is all so difficult to measure and know; this isn't that great an argument.

If Mikovits had a test for XMRV which distinguished between blood samples from CFS patients and healthy controls under blinded conditions, then XMRV's origin would not prove it did not play a causal role in CFS (at least, some CFS). The results from the BWG are a far stronger reason for believing XMRV does not cause, and is not associated with, CFS. If Mikovits aced the Lipkin study (which I'm certainly not expecting), then I think that everyone would need to look again at CFS and XMRV, regardless of the other evidence we have indicating that it does not play a role in CFS and PC.
 
Messages
5,238
Location
Sofa, UK
If you click the link in 'produced in the laboratory'' it states "Because XMRV arose between 1993-1996, this scenario could not explain cases of prostate cancer and chronic fatigue syndrome that arose prior to that date.''

It doesn't say XMRV can't be harmful, but obviously it's impossible to be a cause of ME/CFS or prostate cancer because of these findings and the fact that XMRV is not found in the human population.
Obviously I've heard that argument made before too, and once again it just fries my logic chip; I struggle to believe that a logical person could say something like this in all seriousness. Surely the flaw in this reasoning must be obvious to at least some of the people making that argument. It is obviously not "impossible to be a cause of ME/CFS or prostate cancer" just because this specific viral sequence was created in the lab between 1993 and 1996 and prostate cancer and ME/CFS existed before that time.

It is of course impossible for something created in 1993 to be the cause of cases of prostate cancer and ME/CFS that arose previous to 1993. But single conditions - even well defined ones - can have multiple causes, that is quite clear. And the specific XMRV sequence could be one example of a class of similar pathogens that have existed in the wild and created in labs for decades, and which cause prostate cancer and ME/CFS. And so XMRV might cause these diseases, even if other things cause them too. I mean: obviously.

It's been so long since this discussion that I'd forgotten what it is about it that riles me up so much, but what makes me distrustful and unconfident about the whole business is precisely the illogicality of pretty much every sentence you read from the 'denialist' side of this whole argument. When people keep on making unjustified leaps of logic like...

The retrovirus XMRV does not cause prostate cancer or chronic fatigue syndrome – that hypothesis was disproved by the finding that the virus was produced in the laboratory in the 1990s by passage of a prostate tumor in nude mice.

...I just can't understand why scientists would be so lazy with their words and their logic - and it doesn't engender trust in me when I see so many of them making statements like this, which draw conclusions which quite clearly do not follow from the premise.
 
Messages
5,238
Location
Sofa, UK
I guess that you can argue that if rates of PC and CFS have not dramatically increased since 1993-1996, then it's unlikely that XMRV causes them, or a significant number of them.
Now there's a logically accurate sentence. Why is this degree of accuracy so difficult for scientists commenting on XMRV?
 

ukxmrv

Senior Member
Messages
4,413
Location
London
The CFS cases prior to 1993 could have been caused by another recombination at an earlier time. There is no logical reason why something could only happen once.

There could have been an event in the 1930's which sparked off the ME epidemics. Then another in the 1990's. No one has compared the original ME epidemic patients with the later sporadic cases.
 

FancyMyBlood

Senior Member
Messages
189
Now there's a logically accurate sentence. Why is this degree of accuracy so difficult for scientists commenting on XMRV?

Mark, you made some good points in your earlier post. Racaniello could be more precise, but I'm not sure if you merely made a theoretical fallacy argument (ie. Racaniello's argumentation doesn't support his conclusion) or that it's meant in a more practicaly way (ie. XMRV can cause ME/CFS or PC afterall).

In case you meant the latter, it's now well established XMRV is not a human pathogen, so it's imposssible to be a cause of ME/CFS or prostate cancer. Racaniello obviously knows this (it's what his current blog is all about), so it's not written with bad intentions (I know you didn't claim this).

As a side note, I believe Racaniello is one of the most objective scientist to cover XMRV. He fully acknowledged when he jumped to conclusions to quickly (for example after the Retrovirology papers) and it seems he's one of the only scientists to be interested what caused the serological response.
 
Messages
5,238
Location
Sofa, UK
Mark, you made some good points in your earlier post. Racaniello could be more precise, but I'm not sure if you merely made a theoretical fallacy argument (ie. Racaniello's argumentation doesn't support his conclusion) or that it's meant in a more practicaly way (ie. XMRV can cause ME/CFS or PC afterall).
My point was that the assertion wasn't logically justified. I certainly wouldn't assert that "XMRV can cause ME/CFS or PC" because that is not known. But neither would I assert that it can't, because that is not known.

In case you meant the latter, it's now well established XMRV is not a human pathogen, so it's imposssible to be a cause of ME/CFS or prostate cancer. Racaniello obviously knows this (it's what his current blog is all about), so it's not written with bad intentions (I know you didn't claim this).
I certainly wouldn't claim it was written with bad intentions, that's futile for anyone to speculate about and not relevant to the substantive argument; it's an ad hominem argument and thus a fallacy. But I do not believe it is "well established that XMRV is not a human pathogen", indeed I think that is clearly not established. One of the 3 recent papers stresses dangers to lab workers and talks about a 'biosafety biohazard', so it seems clear that it is still a concern. Just because it hasn't been consistently detected in human blood, and associations with PC and ME have not been confirmed, does not mean it is not a human pathogen. Arguments have been made that it is effectively restricted by APOBEC3 but that is far from conclusive, XMRV also knocks out APOBEC3, some people have genetic deficiencies in APOBEC proteins, and some cells lack that restriction factor. In the first macaque study the virus entered the monkeys' bloodstream then disappeared but showed up later in tissues. So I do not think for one minute that it is scientifically established that XMRV is not a human pathogen, that has merely been 'established' as a narrative for public consumption.

As a side note, I believe Racaniello is one of the most objective scientist to cover XMRV. He fully acknowledged when he jumped to conclusions to quickly (for example after the Retrovirology papers) and it seems he's one of the only scientists to be interested what caused the serological response.
I'm glad you've given me the opportunity to agree with that. I've been impressed by Racaniello's coverage and I was disappointed to have to take issue with his assertion as I did above. I think he generally does a very good job of communicating complex issues to a wider audience without compromise and in an independent-minded way, and that is really important and valuable. He was not quick to jump to unjustified conclusions a year or so ago, at a time when many other commentators were doing so. However, more recently he appears to be doing so, and the quote I've focused on seems to be a case in point. If his conclusions are justified based on other information then he should state in the article what that justification is. I haven't yet seen anything that proves that "XMRV does not cause prostate cancer or chronic fatigue syndrome" much less anything that proves that "XMRV is not a human pathogen". "There is no evidence that..." I could stomach, but definitive statements from scientists about unknowns wind me up.
 
Messages
13,774
There's some phrase like: 'There is no proof in science. Proof is for mathematics and alcohol.'

(Wrote out a long post related to this... realised it was too short to say anything meaningful. Deleted it!)
 
Messages
5,238
Location
Sofa, UK
I get that Esther, but then it's confusing for a mere mathematician when scientists keep saying things like "that hypothesis was disproved". I expect "proved" to mean "proved", and I wish they could use more appropriate words. And I don't think the hypothesis cited was "disproved" even in the scientific sense.
 

FancyMyBlood

Senior Member
Messages
189
I certainly wouldn't claim it was written with bad intentions, that's futile for anyone to speculate about and not relevant to the substantive argument; it's an ad hominem argument and thus a fallacy. But I do not believe it is "well established that XMRV is not a human pathogen", indeed I think that is clearly not established. One of the 3 recent papers stresses dangers to lab workers and talks about a 'biosafety biohazard', so it seems clear that it is still a concern. Just because it hasn't been consistently detected in human blood, and associations with PC and ME have not been confirmed, does not mean it is not a human pathogen. Arguments have been made that it is effectively restricted by APOBEC3 but that is far from conclusive, XMRV also knocks out APOBEC3, some people have genetic deficiencies in APOBEC proteins, and some cells lack that restriction factor. In the first macaque study the virus entered the monkeys' bloodstream then disappeared but showed up later in tissues. So I do not think for one minute that it is scientifically established that XMRV is not a human pathogen, that has merely been 'established' as a narrative for public consumption.

I believe there is overwhelming evidence XMRV is not a human pathogen. Putting aside all the initital negative papers, even the positive papers haven't hold up:
- Singh couldn't find it anymore after she cleared her lab from contamination.
- Hanson isn't able to find MLVs anymore
- Same for Lo and Alter
- Again, same thing for Silverman et al.
- It seems the only one who is able to still find it is Mikovits, but she can't distinguish patients from controls under real blinded conditions. This clearly points to contamination.

Then we have the phylogenetic analyses which doesn't support that XMRV is a genuine human pathogen. The macaque studies are particularly weak to support a role of XMRV in ME/CFS because initital positive papers where able to find XMRV in human blood regardless. If we take this in mind is there any argument that can be made XMRV might be involved afterall? Yes, the serological response might be a straw, but it doesn't seem to be XMRV specific.

If anyone else still has arguments to support XMRV might be a human pathogen, I'm happy to hear them, but I don't think there are many arguments left.
 
Messages
5,238
Location
Sofa, UK
How does the fact that most researchers can't find XMRV in human blood represent "overwhelming evidence XMRV is not a human pathogen"? The point of the macaque study was that it demonstrated XMRV disappearing from blood and taking up residence elsewhere.

Just because the initial positive results found it in human blood does not mean that if we now find that other detections in human blood are likely to be contamination, then the rest of the findings of Lombardi et al are necessarily invalid. The key thing here for me is that just because the initial evidence of human infection has been refuted, does not mean that we can conclude that therefore it's not pathogenic. We can conclude that there's no remaining evidence that it is, but that's not the same thing as concluding that it isn't!

If I said to you that I thought smoking caused cancer by turning blood blue, infecting the lungs, and interfering with genes, and you proved that I was wrong and smoking didn't turn blood blue, would you conclude that this proved that smoking doesn't cause cancer? If you said that cancer existed before people started smoking, would that prove that smoking doesn't cause cancer? Of course not. But those are precisely the same logic as the arguments being said to conclude that "XMRV is not a human pathogen".

The phylogenetic analyses speak to its origins and suggest it isn't mutating as an active communicable infection, but that's not the same as saying it's not pathogenic in humans.

The argument about XMRV specifically is likely a blind alley though, because the question now is really about HGRVs, MLVs and MLVRs in general, and what is the explanation of the multiple results in both PC and ME finding different levels in patient samples vs controls? As one of the 3 studies points out in its extract (and it understates the point), those different levels in the PC findings remain very difficult to explain - and they remain unexplained. Even if there were something in the patients' blood that meant they were more likely to facilitate detection of a contaminant, that could still be a medically relevant clue. For me, that's the part that still needs to be explored and answered; waving it away with "maybe the patient samples were handled more often", as has been done for decades with results like this, isn't good enough; that's a hypothesis and it requires evidence for that to be a scientific proposition.

I think that focusing on XMRV specifically, as defined as a specific genetic sequence, isn't a good plan really - after all, Lo and Alter never detected that specific sequence in their (very roughly) 80% vs 5% results. But I object to the idea that it's not even viable to say that "XMRV might be a human pathogen" - it's a retrovirus, and it replicates in human cells, and there is still no reason I can see to say that it cannot be pathogenic in humans.
 

FancyMyBlood

Senior Member
Messages
189
How does the fact that most researchers can't find XMRV in human blood represent "overwhelming evidence XMRV is not a human pathogen"? The point of the macaque study was that it demonstrated XMRV disappearing from blood and taking up residence elsewhere.
It's illogical to refer to the macaque study because the initital positive papers were able to find XMRV in the blood years after initital 'infection'. It's implausible to assume that the initital positive papers were able to find it in human blood, but subsequent attempts of the same authors weren't because all of a sudden it dissappeared. This would be possible if patients were just exposed to this virus in the initital papers, but they weren't, many already had ME/CFS for years. Lo and Alter tested the same samples more than a decade later and they were still able to find PMLVs. I think Mikovits stated the same for XMRV.

But even if we follow this hypothetical argument it's still impossible that XMRV is human pathogen. It maybe difficult to conclude this from ME/CFS papers because they haven't studied many different tissues (although Natelson did a XMRV spinal fluid study and didn't find anything). However, in PC studies researchers can't find XMRV in prostate tissue anymore. According to your macaque argument they should find it because prostate tissue is the only place XMRV can be present to cause PC. The only possible explanation left is contamination. Very simple.

Just because the initial positive results found it in human blood does not mean that if we now find that other detections in human blood are likely to be contamination, then the rest of the findings of Lombardi et al are necessarily invalid. The key thing here for me is that just because the initial evidence of human infection has been refuted, does not mean that we can conclude that therefore it's not pathogenic. We can conclude that there's no remaining evidence that it is, but that's not the same thing as concluding that it isn't!
Wait a second...... I'm afraid we're arguing two different things. I'm trying to make the point that XMRV is not a human pathogen, simply because it's not circulating in humans. I believe you're trying to say that it still may be pathogenic, even if it's not circulating in humans (ie. it's able to infect human cells and may eventually spread from the laboratory to the human population). Please correct me if I'm misunderstanding it.

The argument about XMRV specifically is likely a blind alley though, because the question now is really about HGRVs, MLVs and MLVRs in general, and what is the explanation of the multiple results in both PC and ME finding different levels in patient samples vs controls? As one of the 3 studies points out in its extract (and it understates the point), those different levels in the PC findings remain very difficult to explain - and they remain unexplained. Even if there were something in the patients' blood that meant they were more likely to facilitate detection of a contaminant, that could still be a medically relevant clue. For me, that's the part that still needs to be explored and answered; waving it away with "maybe the patient samples were handled more often", as has been done for decades with results like this, isn't good enough; that's a hypothesis and it requires evidence for that to be a scientific proposition.
I believe there is sufficient evidence to support that argument (bolding is mine). Mikovits wasn't able to distinguish controls from patients under real blinded conditions (BWG study). She used her own lab and the same techniques as she did in the originial Science paper, so the ONLY explanation is that, indeed, the samples were handled differently.
 

VillageLife

Senior Member
Messages
674
Location
United Kingdom
They were able to find the new human retrovirus in the UK Ashford hospital group and they were tested at the NCI.
And they tested 50 UK blood donors as controls. And a small % of them were positive.
 
Messages
5,238
Location
Sofa, UK
It's illogical to refer to the macaque study because the initital positive papers were able to find XMRV in the blood years after initital 'infection'. It's implausible to assume that the initital positive papers were able to find it in human blood, but subsequent attempts of the same authors weren't because all of a sudden it dissappeared. This would be possible if patients were just exposed to this virus in the initital papers, but they weren't, many already had ME/CFS for years. Lo and Alter tested the same samples more than a decade later and they were still able to find PMLVs. I think Mikovits stated the same for XMRV.
In no way am I assuming that the initial positive papers found it in human blood but then it disappeared. That would obviously be a ridiculous argument and not supported by the evidence. My point about the macaque study is that even if XMRV was not present in the blood, and the original detection was due to contamination, failing to find it in blood does not mean it can't be present in the body.

But even if we follow this hypothetical argument it's still impossible that XMRV is human pathogen.
Again with that word "impossible". This is what I'm objecting to. It is "impossible that XMRV is human pathogen". Why then does one of the three new papers state this conclusion:

Overall, the replication-competent retrovirus XMRV, present in a high number of laboratories, is able to infect human lymphoid tissue and produce infectious viruses, even though they were unable to establish a new infection in fresh tonsillar tissue. Hereby, laboratories working with cell lines producing XMRV should have knowledge and understanding of the potential biological biohazardous risks of this virus.

http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0037415
Why do researchers still worry about "potential biological biohazardous risk" if it's "impossible that XMRV is [a] human pathogen?"

It maybe difficult to conclude this from ME/CFS papers because they haven't studied many different tissues (although Natelson did a XMRV spinal fluid study and didn't find anything). However, in PC studies researchers can't find XMRV in prostate tissue anymore. According to your macaque argument they should find it because prostate tissue is the only place XMRV can be present to cause PC. The only possible explanation left is contamination. Very simple.
Even if it isn't present in prostate tissue and even if it isn't found in spinal fluid in ME/CFS patients, it is still not impossible for XMRV to be pathogenic! Even if contamination takes place in experiments - which I think is a given - that does not make it impossible for the contaminant to be pathogenic.

Wait a second...... I'm afraid we're arguing two different things. I'm trying to make the point that XMRV is not a human pathogen, simply because it's not circulating in humans. I believe you're trying to say that it still may be pathogenic, even if it's not circulating in humans (ie. it's able to infect human cells and may eventually spread from the laboratory to the human population). Please correct me if I'm misunderstanding it.
You're stating that XMRV is not a human pathogen, and concluding from the failure to detect it in human blood or prostate tissue that this means it is definitely not circulating in humans. I'm saying that neither is certain or proven. Just because you can't find your keys in your pocket, on your table, or under your sofa, does not mean that you can claim to have proven they are not in your house and they don't unlock doors anyway.


"maybe the patient samples were handled more often"
I believe there is sufficient evidence to support that argument (bolding is mine). Mikovits wasn't able to distinguish controls from patients under real blinded conditions (BWG study). She used her own lab and the same techniques as she did in the originial Science paper, so the ONLY explanation is that, indeed, the samples were handled differently.

I have seen no evidence that the patient samples were handled more often than the controls in Lombardi et al, nor have I seen any evidence that the prostate cancer cells were handled more often than the controls in Lo/Alter's study or in the prostate cancer studies. In order for this explanation to hold up, the patient samples must have been handled more often in all of these different studies, by many different researchers. Even if this were true, it would raise the question: "why would a good scientist do this, and how can we get scientists to stop making such basic errors?"

Claiming "that is the only explanation" and that this constitutes evidence is completely flawed reasoning, and that sort of argument trips us up all the time when we encounter possibilities that we had not thought of. "That is the only possible explanation" requires you to have constructed a logical exploration of every single possibility, and you have to prove that every single other possible explanation does not fit the facts. Yes, it's true that "once you have eliminated the impossible, what remains, however improbable, must be the truth" - but this argument rests on the assumption that you have considered every single possibility, however improbable.

There are other possible explanations for the facts you cite, however improbable. I am not claiming the following examples as plausible or saying I believe they are true, but they are illustrative of the logical fallacy of claiming something must be true because it's the only explanation you can think of.

Here are some alternative top-of-the-head scenarios in which the original results were valid, that fit the known facts. The BWG samples provided could have been flawed or mixed up in some way. The blinding/coding could have gone wrong - and why is that so much less plausible than the basic error of handling samples differently? There could have been cross-contamination between patient and control samples during the process of transferring them to the WPI. There could be some factor in the reagents in the tubes that hid or 'deactivated' any XMRV present, and thus the only positive signals the WPI could find was then from contamination. The WPI, under time pressure, might not have had time to carry out their work to the same standards as the original experiment. Some essential agent in their detection process may have degraded or been tampered with in the time between the two studies. There could be some factor essential to successful detection which they performed in the initial experiment which the WPI weren't aware of. The samples could have been tampered with en route and the labelling mixed up. All of these possibilities are improbable, and the last one wanders into the realm of conspiracy theory - but none of these scenarios are impossible.

And even assuming that the original experiments by Lombardi et al, Lo/Alter and the prostate cancer researchers were all fundamentally flawed in some way, that still does not mean that the flaw was "more frequent handling of the patient samples". Anomalous results like these have been reported for decades and consigned to the dustbin with a guess of "probably handled more often", but that is no more than a hypothesis, not even a theory, without evidence that this was the case. Maybe the patient samples had different handling - through the reagents in the tubes, the needles used to extract the blood, the length of time they were stored or transported, or some other obscure factor that none of the researchers thought to control for. Another theory that has been advanced for such anomalous results can be dismissed - the idea that researchers conduct many such trials and only publish the successful ones - because the p-values in these cases are way too high to justify that explanation (although this is a genuine major issue and illustrates why all experiments should be required to register before they are conducted, and publish even if their results are unsuccessful). Or maybe these experiments were all fraudulent or self-deceiving; maybe somebody snuck in one night and put XMRV in the patient samples; maybe they all just lied...such things do happen.

My aim here is not to pick one of the hypothesis from the above two lists; I am fairly confident that the true explanation is one that I have not thought of. I'm merely pointing out that it isn't satisfactory to say there is sufficient evidence to support the hypothesis that one set of samples were handled more often than the other, when in fact there is no such evidence. The most your argument could support is the case that there was something wrong with the methodology of all these studies - and many others before them - and I'm saying that if this sort of thing keeps going wrong then it would be more than worthwhile to try to figure out exactly what this common methodological error is so that it can be avoided in the future.

You can't just keep throwing anomalous scientific results in the bin marked "I don't know, something must have gone wrong, maybe we just imagined it" on the basis of arguments based on exclusion...or they'll end up like CFS patients...
 

FancyMyBlood

Senior Member
Messages
189
In no way am I assuming that the initial positive papers found it in human blood but then it disappeared. That would obviously be a ridiculous argument and not supported by the evidence. My point about the macaque study is that even if XMRV was not present in the blood, and the original detection was due to contamination, failing to find it in blood does not mean it can't be present in the body.

Again with that word "impossible". This is what I'm objecting to. It is "impossible that XMRV is human pathogen". Why then does one of the three new papers state this conclusion:

Why do researchers still worry about "potential biological biohazardous risk" if it's "impossible that XMRV is [a] human pathogen?"


Even if it isn't present in prostate tissue and even if it isn't found in spinal fluid in ME/CFS patients, it is still not impossible for XMRV to be pathogenic! Even if contamination takes place in experiments - which I think is a given - that does not make it impossible for the contaminant to be pathogenic.

I'm watching the (boring) Champions League final right now and it's break so I have a limited time to post atm. I'll get back later to the rest of your post.:)

It's impossible to be a human pathogenic if it's not circulating in humans. Yes, it might have the potential to be pathogenic, but it's not spreading in humans. This may be just semantics over the words 'human pathogenic', so again, I'm refering to a pathogen that is circulating in humans and causing disease.

You're stating that XMRV is not a human pathogen, and concluding from the failure to detect it in human blood or prostate tissue that this means it is definitely not circulating in humans. I'm saying that neither is certain or proven. Just because you can't find your keys in your pocket, on your table, or under your sofa, does not mean that you can claim to have proven they are not in your house and they don't unlock doors anyway.
Mark, there have been literally 1000's of samples analyzed all over the world. Not only in PC or ME/CFS but also in totally different groups like immunocompromised patients (for example people with HIV). If XMRV was a circulating pathogenic it would've been found by now. I guess you can always use the argument that it's still not enough and we should screen the whole world but that's just not valuable. But with that same logic I could claim that there is a possibility that there is a human out there with 50% cow DNA. It's just an illogical argument.[/I]
 

FancyMyBlood

Senior Member
Messages
189
I have seen no evidence that the patient samples were handled more often than the controls in Lombardi et al, nor have I seen any evidence that the prostate cancer cells were handled more often than the controls in Lo/Alter's study or in the prostate cancer studies. In order for this explanation to hold up, the patient samples must have been handled more often in all of these different studies, by many different researchers. Even if this were true, it would raise the question: "why would a good scientist do this, and how can we get scientists to stop making such basic errors?"

Claiming "that is the only explanation" and that this constitutes evidence is completely flawed reasoning, and that sort of argument trips us up all the time when we encounter possibilities that we had not thought of. "That is the only possible explanation" requires you to have constructed a logical exploration of every single possibility, and you have to prove that every single other possible explanation does not fit the facts. Yes, it's true that "once you have eliminated the impossible, what remains, however improbable, must be the truth" - but this argument rests on the assumption that you have considered every single possibility, however improbable.

There are other possible explanations for the facts you cite, however improbable. I am not claiming the following examples as plausible or saying I believe they are true, but they are illustrative of the logical fallacy of claiming something must be true because it's the only explanation you can think of.

Here are some alternative top-of-the-head scenarios in which the original results were valid, that fit the known facts. The BWG samples provided could have been flawed or mixed up in some way. The blinding/coding could have gone wrong - and why is that so much less plausible than the basic error of handling samples differently? There could have been cross-contamination between patient and control samples during the process of transferring them to the WPI. There could be some factor in the reagents in the tubes that hid or 'deactivated' any XMRV present, and thus the only positive signals the WPI could find was then from contamination. The WPI, under time pressure, might not have had time to carry out their work to the same standards as the original experiment. Some essential agent in their detection process may have degraded or been tampered with in the time between the two studies. There could be some factor essential to successful detection which they performed in the initial experiment which the WPI weren't aware of. The samples could have been tampered with en route and the labelling mixed up. All of these possibilities are improbable, and the last one wanders into the realm of conspiracy theory - but none of these scenarios are impossible.

And even assuming that the original experiments by Lombardi et al, Lo/Alter and the prostate cancer researchers were all fundamentally flawed in some way, that still does not mean that the flaw was "more frequent handling of the patient samples". Anomalous results like these have been reported for decades and consigned to the dustbin with a guess of "probably handled more often", but that is no more than a hypothesis, not even a theory, without evidence that this was the case. Maybe the patient samples had different handling - through the reagents in the tubes, the needles used to extract the blood, the length of time they were stored or transported, or some other obscure factor that none of the researchers thought to control for. Another theory that has been advanced for such anomalous results can be dismissed - the idea that researchers conduct many such trials and only publish the successful ones - because the p-values in these cases are way too high to justify that explanation (although this is a genuine major issue and illustrates why all experiments should be required to register before they are conducted, and publish even if their results are unsuccessful). Or maybe these experiments were all fraudulent or self-deceiving; maybe somebody snuck in one night and put XMRV in the patient samples; maybe they all just lied...such things do happen.

My aim here is not to pick one of the hypothesis from the above two lists; I am fairly confident that the true explanation is one that I have not thought of. I'm merely pointing out that it isn't satisfactory to say there is sufficient evidence to support the hypothesis that one set of samples were handled more often than the other, when in fact there is no such evidence. The most your argument could support is the case that there was something wrong with the methodology of all these studies - and many others before them - and I'm saying that if this sort of thing keeps going wrong then it would be more than worthwhile to try to figure out exactly what this common methodological error is so that it can be avoided in the future.

You can't just keep throwing anomalous scientific results in the bin marked "I don't know, something must have gone wrong, maybe we just imagined it" on the basis of arguments based on exclusion...or they'll end up like CFS patients...

Mark, I don't believe I'm comitting a logical fallacy. None of your 'improbable' other explanations make any sense. Many have been debunked already in the BWG FAQ and the rest is just grasping at straws. Anyone could come up with 1000 other possible explanations. I mean maybe a vaccine company kidnapped one of the delivery guys and spiked XMRV at the same amount in all of the WPI samples. Heck, I could even go further and claim not HIV, but XMRV might be the cause of AIDS because 'insert 500 pages of possible explanations'.

It's interesting to follow your argument though, let's assume Mikovits was able to distinguish patients from controls in the BWG study. Anyone could claim that it still doesn't tell anything by using your same 'improbable other explanations'. I mean maybe their sample delivery guy was kidnapped again, but now by a big pharm company that produces anti-retrovirals. The list with other 'improbable other explanations' goes on and on...... Heck, no paper means anything anymore because there are countless 'improbable other explanations'.

I guess I'll never convince you, but there is overwhelming evidence to support that XMRV is a contaminant. A couple of months ago I would agree with you about the sampling issue. However, the BWG results were the final nail in the coffin for me. Even Mikovits hasn't disputed those results and I believe that says enough.

It may be still interesting to know why/how the samples were handled differently, but we'll probably never know the reason. In the worst case scenario, scientist aren't going to admit they commited fraud. In the scenario they were 'just' sloppy they're not going to commit it either.