Check this out:
From what I gathered, the original published results seemed to fly in the face of what had been accepted wisdom, based on a fairly substantial body of evidence. In and of itself, that is not necessarily a problem. Sometimes someone comes up with a novel way to examine a research question and finds something useful. Needless to say, Brockman and colleagues attempted to replicate the LaCouris and Green findings and instead found some glaring anomalies that suggested potential fraud. There's an old saying in the business world: "trust but verify." That is true in the sciences as well. Sometimes that verification process breaks down, and when it does, it opens up the risk for very regrettable outcomes. For those pushing for open archiving of published data and the like, I think this will give some needed fuel for the fire.
In the meantime, we're left with another truism: "If it seems to good to be true, it probably is."