On October 20, the Journal of the American Chemical Society retracted a 2009 paper. The retraction notice for “Single Gold Nanoparticles Counter: An Ultrasensitive Detection Platform for One-Step Homogeneous Immunoassays and DNA Hybridization Assays” was somewhat opaque:
This article is being retracted due to inaccurate DNA hybridization detection results caused by application of an incorrect data processing method. The authors regret any confusion that may have been created by the paper’s publication.
We contacted the paper’s lead author, Jicun Ren, of the College of Chemistry and Chemical Engineering, State Key Laboratory of Metal Matrix Composites, Shanghai Jiaotong University, to ask for more detail. He responded:
The method and principle of a single gold nanoparticle counting technique is correct. The result of immunoassay is also right. The problem is that the limits of detection in DNA hybridization assay is not 1 fM, and is about 1 pM. In our paper, the photon burst counts were obtained by the “pick peaks” tool in Origin software using 3 times ratios of signal to noise. In data processing about DNA hybridization detection, we incorrectly decided the noise and background signal values, which led to this wrong result.
We regret any confusion that may have been created by the paper’s publication.
Ren’s response seemed transparent, but that doesn’t mean we understood what it means. For that, we checked in with our molecular biology guru Jeff Perkel. He tells us that Ren and his colleagues designed the technique to detect flashes of light thrown off by gold nanoparticles in order to measure their concentration. But it seems they forgot to correct for something, and ended up with a technique that looked about three orders of magnitude more powerful than it actually is.
The whole point of the paper was that the technique was two to five times more powerful than existing methods, so once you take away that inaccurate extra power, you’re left with a technique that only works as well as the existing one. It made the findings a bit “so what?”
The paper has been cited 14 times, according to Thomson Scientific’s Web of Knowledge. We tried contacting the journal’s editor, but haven’t heard back.