Warning: This blog contains stories about phantom serial killers, frankenfoods, mysteriously phosphorylated bands and unrequited ligations that may be disturbing to some people. Children or scientists prone to anxiety over irreproducible results should read this with their eyes shut.
Clouds hung low in the sky, and the late October wind howled between the buildings, rattling the window panes of the basement laboratory. The grackles cawed in desperate warning, their flocks changing the evening color palette from gray to black. I was as unsettled as the weather, watching my blot slosh back and forth.
For two months I had pursued this exciting new target that was phosphorylated in cells when activated by our lab’s protein of interest. I returned to the lab at all hours to passage my cells at precisely the right moment, single-minded in my pursuit of this new protein band that I had found—in my first weeks working on my very first project in my new graduate lab. It seemed too good to be true. When I presented at the group meeting tomorrow, everyone would hail me as “genius” and “science prodigy”.
My presentation should have been done by now, but the senior post doc in the lab asked if I had tested my antibodies, primary and secondary, to see if they reacted with the protein preparation used to treat my cells. “Silly”, I thought, “these are the same primary and secondary antibodies everyone else in the lab uses. Nobody else has problems.” Still I knew he would ask that same question again tomorrow, and my advisor would insist on doing the control experiments. So, here I was, late again in the lab, doing yet another Western blot to check the antibodies for bad behavior.
Soon I was ready to look at my blots, and finish my lab presentation. “What was this? The secondary-no-primary control is reacting with the treatment extract?”
Mumbling a few choice words that cannot be recorded here, I opened my laptop and deleted all of the slides from my lab presentation. Group meeting will be very short tomorrow.
“Frankenfoods” was the overly-dramatic name given to food stuffs derived from genetically modified organisms when GMO foods first hit the market; it was a misnomer completely lacking any scientific credibility. But it was sensational, so it stuck.
I was working on developing methods to detect the GMO plants, seeds and foods for research purposes. Transgenic crops use the promoter from the cauliflower mosaic virus (CaMV) to drive the production of whatever gene has been introduced into the crop, whether it’s a gene that allows the crop to produce high quantities of a vitamin, grow in drought conditions, or confer herbicide resistance.
I used commercially available primers for the CaMV 35S promoter, which is used in transgenic crops, in my assay to test some seeds that had been previously verified as non-GMO. To my surprise, these seeds tested positive for product amplified by the primers, which I interpreted as meaning the seeds were either improperly labeled or somewhere along the line they had mixed with GMO seeds.
But that was a mighty bold claim, with some serious implications, so I thought about how I could verify my own results. Running a few controls changed my interpretation. The negative control to test for the presence of CaMV coding sequence that is present in infected plants but not the transgenic promoter produced product. This indicated that these seeds had come from plants that were infected with CaMV, a perfectly natural occurrence. A more rigorous primer set that would amplify product only when the CaMV 35S promoter was attached to the transgene I was interested in did not generate product. These seeds were indeed true to label, non-GMO. Now when I teach workshops on GMO testing, I include this story in the training. Controls are essential to the correct interpretation of results.
From 1993 to 2009, a mysterious serial killer was stalking Europe—Germany, Austria and France—committing brazen thefts, heinous crimes and horrific murders. DNA evidence consistently produced a profile of a single Eastern European female in 40 different crimes, who became known as the Phantom of Heilbronn (because of a murder in that city to which she was linked). She remained an enigma: never appearing on security camera tapes, described by some witnesses as a male, and her very existence denied by accessories who had been arrested in the some cases, even when they were presented with the DNA evidence.
For eight years, the Phantom escaped detection—except for her DNA which turned up on surfaces as diverse as coffee cups, kitchen drawers and heroin syringes . Even after 2 million Euros and over 16,000 person hours spent searching for her, authorities had no leads. The investigation involved more than 400,000 police officers in Germany and Austria. Finally, in 2009 the big break came. The Phantom’s DNA turned up in the sample to identify the burned body of a male asylum-seeker in France, which raised red flags, since the Phantom DNA was female, and the burned body was male.
Now the possibility of a DNA artifact seemed highly likely. Running negative controls of the swabs used to collect the DNA samples at the crime scenes by the various agencies revealed that the swabs themselves contained the DNA of the Phantom of Heilbronn, a production worker in the Eastern European factory that produced the swabs.
But for a simple negative control to test for contamination, millions of dollars and thousands of hours could have been saved, and cases that were cold and unsolved might have been closed.
Okay, there is no story about an unrequited ligation here, but it sounds good. We do however hope these stories illustrate the importance of running the proper controls with your experiments. At Promega, we make a habit of suggesting controls in our technical literature to help you rule out artifacts and troubleshoot confusing results. Still, if you have data that just don’t make sense, give our technical services scientists a shout. They are amazing when it comes to getting runaway data under control.
Have your own lack-of-control science horror story? Share it with us!
Latest posts by Michele Arduengo (see all)
- World Art Day: Experiences of Fine Art That Disrupt and Drive Innovation - April 15, 2022
- Using Structural Computation Models to Predict Productive PROTAC Ternary Complexes - March 7, 2022
- Run to Remember: A Mouse-Model Study Investigating the Mechanism of Exercise-Induced Neuroprotection - January 5, 2022