This happened over the holidays, and I totally missed it:
Han apparently added human blood components to the rabbit blood to skew the results. The human blood came from people whose bodies had produced antibodies to HIV, the virus that causes AIDS, Bradac said. The presence of these antibodies in the rabbits’ blood made it appear that the vaccine was spurring the animals to build defenses against HIV. “This positive result was striking, and it caught everybody’s attention,” Bradac said.
I have to echo James Bradac, here:
“When we saw these exciting developments, we didn’t at that time think there was any problem with contamination,” Bradac said. “Once someone comes out with very interesting result, they are asked by a lot of people to share their re-agents. It’s difficult to pull this off and it’s difficult not to be detected. This went on for several years and it wasn’t detected until January 2013.”
If collaborators (or competitors) suspect samples have been spiked with something, and they think they know what that something is, it is so easy to figure it all out. For example, when Robert Silvermans lab suspected something was 'up' with the CFS patient samples he got from the WPI, they looked for things that should not have been in the patient samples: antibiotic resistance genes and promoter-virus junctions. These are things one would find in an infectious molecular clone (plasmid), but something one would never find in humans.
Likewise, it would have been fantastically trivial to figure out if samples from 'rabbits vaccinated against HIV' made their own anti-HIV antibodies, or the samples had been spiked with human anti-HIV antibodies. Rabbit antibodies are not identical to human antibodies. This fact has some scientific pros and cons (sometimes its obnoxious for researchers, sometimes its convenient), but it is a fact.
I do not know how Dong-Pyou Han got caught, but here is how I would have determined whether the vaccinated rabbit samples were spiked with human anti-HIV antibodies:
1-- Run HIV proteins out in an acrylamide gel (gel electrophoresis). The HIV proteins will be 'invisible', but they are there.
2-- Transfer the proteins from the gel to a piece of nitrocellulose (Western Blot).
3-- Mix the paper+HIV proteins stuck to it with sera from the vaccinated rabbits (primary antibody).
4-- Use anti-human antibodies for my secondary antibody.
IF Han was innocent, the Western Blot would look like this:
The HIV proteins would be on the paper. And the rabbit anti-HIV antibodies would stick to the HIV proteins. But since anti-human secondary antibodies wouldnt 'see' the rabbit antibodies, there would be no way to visualize the proteins. And, since there were no human anti-HIV antibodies, there was nothing for the anti-human secondary antibody to stick to. So, the gel looks empty.
IF Han were guilty, and there were human anti-HIV antibodies in the sera, those antibodies would stick to the HIV proteins on the gel. Then, the secondary anti-human antibodies would stick to the human antibodies, and the HIV protein bands would appear:
Again, I dont know how Han was ultimately found out, but it would take one work day to do this experiment and figure everything out. Once someone had suspicions, it would have been trivial to resolve this.
It just baffles me why someone would ever pull this kind of stunt, especially when the findings are so high profile (a protective HIV vaccine), and many others will be trying to replicate the results. And its not like negative results are career ending.
Oh my god you have an HIV vaccine that doesnt work!?!?!
So does everyone else.
You can at least publish your negative results so others dont waste their time on it. You reformulate, rethink, and try again.
Baffled why this happened.
Glad he was found out so more $$$ wasnt wasted.
I wonder if this is a case of "publish or perish" gone to extremes, or fear of failing academically? Not sure why someone would take the risk of losing all academic credibility when discovery would be so easy.
Either way, it's incidents like this that have bad consequences to future funding and research in general.
Before one checked for human anti-body spike in, one would have to suspect it. Maybe that's why I'm not-so-baffled.
Rork, exactly. If you have the tainted samples and decided something was fishy, "he must have spiked it with human blood" is not necessarily the first thing that comes to mind. Testing that hypothesis is easy (as Abby shows), but deciding that is the thing to test is not the first obvious deduction.