Experiments gone wrong

I'm editor of the Lab Tools section here at The Scientist. The idea behind Lab Tools is to root out aspects of lab techniques that give researchers a headache, and try to offer some tips and tricks that people have used to troubleshoot them.

My position here seems appropriate, because my grad school research was derailed when a biochemical assay which formed the core of my dissertation suddenly stopped working. I replaced every reagent and piece of equipment, but never figured out what went wrong. I did gripe in the corridors -- a lot -- and found that many of my colleagues had similar sob stories. Those stories always made my day – not just because it was nice to know that my work wasn’t more ill-fated than others, but because it reminded me how much of an art bench work can be.

Do you have a story of an experiment gone mysteriously wrong? Did yours end better than mine?

In 1957 I was a PhD student researching the transmission of the mosquito-borne Semliki forest & West Nile viruses at the London School of Hygiene & Tropical Medicine. I attempted to produce primary mosquito cell tissue cultures by adapting the existing technique for making primary chick embryo cell cultures. It seemed obvious to me that fetal calf serum (which worked for chickens) would not work for insects, and that instead of bovine serum albumin I needed a supply of insect serum, i.e. hemolymph. I used macerated colonized Aedes aegypti larvae as source of cells and the filtered fluid as an analogue for fetal calf serum, an artificial insect hemolymph concoted from expensive amino-acids and a buffer solution, and incubated the mix at room temperature. It didn't work. Much was my chagrin, a few years later, to learn that somebody had succeeded in culturing mosquito cells using fetal calf serum & bovine serum albumin. I had been far too logical!

My experience probably wouldn't happen today because most molecular work is done using kits but back in the 90s we were doing Northern blots (mRNA gels to blots and probed with cDNA). One day they stopped working. We went through every single component one by one and never found the answer. The process took almost a year and needless to say we were more than frustrated. Then one day it dawned on my coworker that Maniatis suggested that RNA should be precipitated (if I remember correctly) using an ammonium chloride buffer which is what we had done to the current samples while normally DNA is precipitated using a sodium chloride buffer. After re-precipitating the RNA with the DNA buffer the Northerns worked! Turns out the ammonium chloride interfered with the hybridization. Who knew?

My masters' thesis experiment has been plagued with problems since its beginning.

1) The training protocol to get naive pigeons to key-peck for food reinforcement didn't work, despite being the same protocol (and computer program) previously and successfully used to train 24 other pigeons. After three weeks tweaking a protocol that should have worked in one day I ended up shaping the key-peck response for each pigeon individually, much in the way you would teach a dog to sit, which took three hours total for all the subjects.
Time lost: 3 weeks

2) The installation of our lab's new ventilation system was running behind schedule resulting in loud drilling through concrete, banging on giant metal ducts, and all around bedlam which delayed the experiment just after training.
Time lost: 2 weeks

3) Thirty days into the experiment proper, and injections of saline and two low doses of amphetamine, the third dose (a moderate dose) evoked regurgitation for every single subject. This was something that I had never encountered in several experiments using amphetamine on pigeons (not to mention drug from that exact bottle). My adviser who had done many experiments with amphetamine and pigeons never encountered it. We polled colleagues who account for a good portion of published "pigeons on amphetamine" papers and nobody had experienced this. Theories du jour were coming in from our consulting vet and colleagues. The pigeons have since become tolerant to the bizarre emetic effect, though the experiment had to start over.
Time lost: 40 days

On the good side, the data are beautiful and things are going quite well now. The project should be complete and defended by the end of August, only 4 months behind the expected completion date. It's also beneficial to have a good adviser who understands these kinds of issues and allows you to take your doctoral candidacy qualifications despite not completing the masters' thesis, which keeps your overall graduate school career on track.

Intersting post guys.. cheers..

I totally agree with Michael. Young scientists should now their future.

Two years of my doctoral research went up in smoke when the forest that comprised my study site burned. Ouch!

LisaICN000309924 wrote:Two years of my doctoral research went up in smoke when the forest that comprised my study site burned. Ouch!

What a tragic expeience. I had the same damn thing with my harddisk.