0
have you ever encountered something in your work that was just "impossible" to be happening?
Like samples running positive when they missed the key ingredient, really unprobable things hapenning together, poltergeits in a lab screwing the samples mysteriously (
), or just anything you find just sooo strange.If you found an explanation for this "impossible" thing later, share it also.
---------------------------------------------------------------------------------------------------------------------------------------------------------------
I will start. (sorry it's a pretty long story, but I thing it's worth it.. hope it's worth it.. you can't tell these stories at home, for obvious reasons
)I'm optimising two melting curve genotyping assays. If you don't know the method, it uses two adjacent hybridisation probes, like this (on left, Simple probe has a different principle).
Basically one of the probes stays attached (anchor) and the other dissociates depending on the complementarity (sensor), mismatched nucleotide causes earlier dissociation during the slow temperature increase. The mutation you want to detect lies within the sensor sequence. You can see in the graph the genotypes that differ in Tm of the sensor. In my case one probe is 3' Fluorescein, that get's excited and the other one is 5' LC640 that you see emitting in the 640 channel. The energy gets transfered through FRET.
Now obviously you can get signal only when these two are together, because otherwise only fluorescein probe is emitting, in a different channel.
As I said I had two completely different assays on different genes, but was optimising them together. I told my technician all the details and gave her a design scheme for both assays. There were names of the primers and probes and so.
The test run on two samples was positive in only one of the genes, the other one failed I didn't know why, but even the positive result was very very ugly and had a low fluorescence levels. Anyway it was distinguishable, so I desided to run all 10 samples I had, hoping to find a heterozygote or mutant (I only had patients with high risk of this polymorphism, no positive control) to get the second peak, and then continue to optimise on the samples with different genotypes only, to save the reagents and templates. One sample of 10 showed the heterozygote pattern, so I choose this one and one wild-type for further tests.
Then I got to work on this myself on Saturday because the technician left for vacation, I found out, that she took anchor probe from one assay and sensor probe from the other assay, so the appropriate pairs in reactions were unmatched. So in each reaction, there were two fluorescent probes, that binded different genes and therefore they couldn't ever get near each other to transfer the energy and give a signal.
I was happy to find out where the problem was and prepared the reactions the right way. I got nice results on both assays, this time the fluorescence was all right and everything. Just one weird thing, the sample heterozygous in the previous bad run, was now wild-type and not heterozygote.
I went back to analysing what actually was in each reaction in the previous failed experiments, I thought originally that in the "positive" assay, there was the corresponding sensor probe, and from some strange reason it did emit fluorescence on it's own and detected the polymorphism in question. But I found out that in the reaction 1) there was actually the corresponding anchor probe 2) the anchor probe has fluorescein marker. Huh?
So how can fluorescein probe be visible in the 640 channel? The answer to this is probably more simple, there is some part of fluorescein spectra overlaping to the 640 channel. Since there was no other signal, this very low, almost background showed as a distinguishable peaks. OK...
But how can an anchor probe (that lies adjacent to the place of polymorphism and should contain no mutations) show heterozygote pattern? I looked into the sequence of the gene in question and did indeed found an intronic polymorphysm within the sequence of anchor probe. That's why the sample with the right probes showed as wild-type, it was heterozygous in a different polymorphism.
So what happend is due to the mixed up reactions, I detected a polymorphism different than was the assay designed to find
Now, isn't this impossible?
(if you ask whether this would cause potential problem with other samples having this intronic polymorphism, then it probably wouldn't, the sensor probe is designed to have lower Tm than anchor [which is called anchor logically because it stays there, while the other dissociates] and the Tm pattern for the right pair of probes is much lower that any variation in binding of the anchor)
And by the way, the name of the topic "Impossible astronaut" comes from a great sci-fi series Doctor Who, if you love science-fiction and fun, you shouldn't mis it!
