If Ive said it once, Ive said it a thousand times here on ERV– Scientists are wrong all the time. ALLLLLLLL the time. Its what we *do*. We make a hypothesis, design an experiment, collect data, and refine the hypothesis, because the original hypothesis was wrong.
Over and over and over and over and over. Less wrong to slightly less wrong, to slightly less wrong. But we are wrong all over the place.
Sometimes things go really wrong and we go from less wrong to more wrong.
Ideally one realizes this has occured before they publish their experiments/data/conclusions in a paper, but sometimes they dont, and papers have to be retracted. I think the general public might think retraction is a bigger deal than it really is, thanks to high-profile retractions in the case of fraud, but fraud and making a mistake are two vastly different creatures. Retraction after you realized you made a mistake is the right thing to do, as it contributes to the progress of science and prevents others from wasting time.
So that is what happened with Robert Silverman and his lab, and their contribution to the 2009 ‘XMRV–>CFS’ paper in Science:
In our 23 October 2009 Report, “Detection of an infectious retrovirus, XMRV, in blood cells of patients with chronic fatigue syndrome,” two of the coauthors, Silverman and Das Gupta, analyzed DNA samples from chronic fatigue syndrome (CFS) patients and healthy controls. A reexamination by Silverman and Das Gupta of the samples they used shows that some of the CFS peripheral blood mononuclear cell (PBMC) DNA preparations are contaminated with XMRV plasmid DNA.
It sucks, as in its pretty embarrassing, but other scientists cant laugh too hard at Silverman because due to the nature of our work, it could very well happen to them the next day. I literally dont know anyone that looks down on Silverman for what happened. I dont know of anyone Silverman attacked or degraded when they couldnt replicate his data. He just do what he do, like any other normal scientist, so the consensus opinion is pretty much what Myra McClure voiced herself:
Myra McClure of Imperial College London conducted a study in the UK last year that found no trace of the virus in people with CFS – one of almost a dozen to have done so since 2009. She applauds Silverman for the retraction. “Bob Silverman is a good scientist and an honourable one,” she says. “I guess he can only retract the figures that he contributed to the original paper.”
Actually, McClure also mentioned a second thing there, that needs to be discussed…
See, Silverman didnt just say “Hey, we messed up these figures. Im sorry. Here is what happened. But we stand behind the conclusions in this paper.”
According to the WSJ, Silverman took his lab off the ‘XMRV–>CFS’ paper entirely. He cannot take down the entire paper– he can only take his name/his lab off the paper.
To understand why that happened, what caused him to change his mind, we need to look at the science Silverman did to determine he was wrong–
Long-time readers of ERV and/or followers of the XMRV fiasco know that of the many criticisms of the XMRV work, one was that all of the sequences the Whittemore Peterson Institute uploaded to Genbank were identical.
The WPI/Judys response to others pointing that out to her were:
1– We promise for reals its not VP62!!
2– There is, liek, TOTALLY a lot of sequence diversity! Theyre not identical! LOOK AT ALL THIS DIVERSITY!!!!!!
Problem #1 is Silverman sent the WPI VP62 plasmids in 2007 (before they ‘found’ XMRV in patient samples.
Problem #2 is that there is ZERO GODDAMN DIVERSITY in the sequences they uploaded to Genbank. YOU CAN SEE IT WITH YOUR OWN EYES.
But Mikovits was above those petty facts.
Luckily, Silverman wasnt.
To investigate whether the samples he got from the WPI were contaminated with VP62 plasmid, he looked for two things:
1– The antibiotic resistance gene found in VP62. We grow more plasmids in bacteria. To select for the bacteria that have the plasmid we want, we give the plasmid an antibiotic resistance gene. You, nor any virus, let alone a retrovirus, dont contain any antibiotic resistance genes. If this gene is present, it indicated that plasmid contaminated the samples.
2– The Promoter-XMRV junction. Lets say I want to grow up some HIV-1 from an infectious molecular clone. I chemically treat my plasmid, plop it onto some mammalian cells, and in a couple of days, there is TONS of HIV in the cell supernatant. That works because there is an artificial promoter in front of the HIV genome in my plasmid, in my case, the promoter from cytomegalovirus. CMVpromoter–>HIV genome does not exist in real life, anywhere. Its a lab trick. Likewise, CMVpromoter–>XMRV does not exist in real life. If you put one primer in CMV, and one primer in XMRV, and you get a product, that product indicats that plasmid contaminated the samples.
Silvermans lab found VP62 plasmid in their samples. Which means the positives Silvermans lab reported in 2009 were all because of plasmid contamination. The positive results were NOT the result of real infection.
Again, not *that* big of a deal. Plasmids contaminate PCR reactions all the time. Coulda happened to anybody.
But… thats not really the end of it…
See, when you contaminate your reactions with plasmids, its either A) sporadic, or B) in every reaction. If plasmid was just ‘around’, you would have expected to see ~3 positives in the 15 CFS samples, and ~3 positives in the healthy controls. If plasmid got into, say, your molecular grade H2O, you would expect to have all samples positive, even your ‘H2O only’ control reaction.
That is not what happened.
Silverman only found VP62 in the CFS samples. Not the healthy controls.
And they detailed in this paper the *extreme* lengths they went to to avoid plasmid contamination. You know what I do to prevent contamination? I set up my PCR in one room, and add my template in the lab. Thats it. You know what they did?
In early 2009 we received PBMC DNA samples from CFS patients and healthy controls from the Whittemore Peterson Institute (WPI), Reno, Nevada. The PBMC DNA samples were taken directly to a clean room upon arrival and stored in a -20oC freezer in the same room. Precautions were taken to minimize the possibility of cross-contamination of the human samples with laboratory sources of XMRV DNA. In particular, neither plasmid XMRV VP62/pcDNA3.1(-) nor XMRV PCR products were ever taken into the clean room. Also, new pipetmans (Gilson) were purchased for exclusive use in the clean room and were never used elsewhere. At the entrance to the clean room there is a sticky pad on the floor and lab personnel must change lab coats upon entering and exiting the clean room. The clean room is locked when not in use. The PCR reaction mixtures that contained PBMC DNA were pipetted in an AirClean 600 PCR Work Station (ISC Bio Express) in the clean room. The PCR Work Station was purchased for use in the clean room and never used elsewhere. The single-round PCR on human DNA samples was performed in a BioRad PCR thermocycler, used exclusively for that purpose, in a separate room from the clean room. The PCR on the plasmid XMRV VP62/pcDNA3.1(-) was performed in yet another room in a different PCR thermocycler from the one used on patient DNA samples.
Uuuuuum… yeeeeeeah… I dont do that. And Ive never gotten contamination. They go super-neurotic and do get contamination?
Well, that would be possible if the source of the contaminant was one of the previously proposed mechanisms– Mouse DNA is in PCR and reverse transcription reagents, its in Qiagen columns. MLVs of any stripe contaminate all kinds of cell lines as proviruses and exogenous viruses. These mechanisms would cause widespread contamination not matter how hard you tried to prevent it.
… But thats not what Silverman found.
He didnt find mouse ERVs.
He didnt find the same exogenous viral sequence over and over.
He didnt find the same XMRV provirus in every sample because of contaminating cell line DNA.
He unquestionably found VP62 plasmid in the samples he got from the WPI… and only in the CFS patient samples.
Meanwhile, at the WPI, they say they get FANTASTIC results with their assays. The 67% positive rate flew up to, what, 100% after the Science publication…
And yet, when WPI/Mikovits are given samples where they do not know beforehand who is ‘supposed’ to be positive and who is ‘supposed’ to be negative, they cannot differentiate between CFS/Healthy/Positive controls. 50:50, implying that half of the people they say are positive are really negative, and half of the people they say or negative are positive, or in other words, they have no idea what they are doing.
When samples are collected from ‘XMRV positive’ patients without any ‘processing’ at the WPI, the samples come up negative.
And yet the CFS samples shipped to Bob Silverman in 2009 were contaminated with XMRV PLASMID before his lab touched them, after WPI touched them, after Silverman gave them the VP62 plasmid.
Isnt that weird?
I think thats just weird.
I mean, how would you get weird results like that?
Thats just so weird!
Well, I guess Silverman thought that was weird too, so he took his lab off the paper.
And lets just say, I dont think he did anything wrong. I think he did what any scientist would do, were they in his position. I think he actively recognized that something was ‘weird’ and investigated it, determined the root cause, and corrected the field.
That is how you make a mistake and correct it gracefully in science.
Quoting Jon Cohen and Martin Enserinks excellent article again:
Back in Reno, Mikovits and Lombardi began feeling besieged. “After the first negative study, it was a dog pile,” Lombardi says. “Let’s be honest: A number of people in the mainstream medical community heard chronic fatigue, and they rolled their eyes and laughed.”
The problem, as they saw it, was that nobody was following their recipe exactly.
Hmm. Yes, it does seem that others are missing a key ingredient to the WPIs ‘recipe for success’.