Question For My Readers

May 22, 2009

One of the things I have always loved about being a scientist is troubleshooting experimental protocols. Anyone with coolio motherfucking troubleshooting stories, please share in the comments. This applies to science troubleshooting as well as other kinds of troubleshooting in non-scientific arenas.

16 Responses to “Question For My Readers”

  1. Blech! I hate troubleshooting. I can do it, I’m reasonably good at it, but really I just want stuff to work like it’s fucking supposed to without any coddling from me. Next time I have something that needs trouble-shooting I’m sending it to you. You can have fun with it and I can spend my time that I would otherwise be faffing around with parameters and shit, getting actual data. Win-win.

  2. You want coolio motherfucking stories or motherfucking horror stories? I’ve got a ton of the latter …

    How about the time I spent 12 months collecting all the samples for my PhD thesis and then couldn’t figure out why none of my subsequent assays would work … I’d put a decimal point in the wrong place and diluted ALL of the samples by a factor of 10 greater than was required. Yeah – spent several weeks working out how to fix that monster-sized fuckup.

    Or what about the time I spent 6 months working up my first goddamned assay* on my own** and couldn’t figure out why the absorbances for all of my standards and samples was maximal … the reaction involved nitrite and it turned out the fucking glass tubes had been sprayed with nitrite during the manufacturing process. Needless to say I totally lost my junk when I found that out, stormed out of the lab and cried for a week.

    Phew … glad to be stepping back from the bench I can tell you.

    * it was pivotal for my work
    ** I was the only person in Grad Dept doing anything remotely close to bench work

  3. MitoScientist Says:

    When working on a joint project with a major pharmaceutical company, we couldn’t get any of our assays to work on the cardiomyocytes they sent. We were supposed to be doing experiments on human cells, and that’s what our antibodies were targeted against. However, we had some monoclonals that also could cross-react with rat samples. We got signals with those assays, and subsequently tried out rat-specific antibodies. Turns out that all the human cardiomyocytes had been contaminated with rat cells, which had then overgrown the entire sample. Pretty dicey if you ask me! Although our lab thought it was hilarious, outside of setting the project back a bit.

  4. Geeka Says:

    When I was just starting out in grad school, one of the tasks that I had to do was immunofluorescence on a transfected population of cells. I couldn’t get it to work. When I talked to my advisor, he instructed me that it was always the fixation or the blocking. So I read, and read, and repeated things. I constructed a 3×3 foot flow chart of IF. Over 6 months I had systematically done ever possible combination of fixing/blocking/staining that one could do. I used different reagents, times, temps…everything. When it came down to the answer…well, advisor was having me stain for something that wasn’t going to work (was physically impossible). I didn’t lose my shit. I just looked at him and asked what we needed to do to fix it. The stains ended up being nice controls.

    Also, I now can look at someone’s IF and immediately tell them what they did wrong.

  5. jc Says:

    I hit the motherload of fuckups during my PhD work. I had preshus samples that made absolutely no fucking sense whatsoever. After 3 more trips to get more of the preshus, same shit! I ran, reran, other people ran, they could run themselves… same shit.

    I hung up the strange results all over my office, hoping for divine inspiration. Every invited speaker who would sit down in my office would get grilled about what s/he thinks is the *problem* – everyone stared blankly (or told me to rerun them for the eleventy time).

    My advisor comes prancing in one day and says “what about this?” and slaps down a paper on my keyboard. I swear to gawd, my ass has never moved so fast to the library in my life. BINGO!

    It turns out that a new guy down the hall was fighting with the same problems for a totally different project but not telling anyone. He went into my advisor’s office wondering WTF is up with HIS dataset… same exact shit!

    We all have a PNAS paper. W00T!!11!

  6. My entire Ph.D. project involved studies performed on a monolayer of cells. Growing the monolayer, however, was a big problem and took me maybe 9 months to figure out. As it turns out (never having worked with mammalian cells before), when I was counting the cells under the microscope to do the seeding, I was counting all of the specks and debris and non-cell shit as cells, and so my cell counts were too low. One day, I was looking under the microscope, and my increasingly-intelligent-about-cells-self said, “Self? Do you really think those specks are cells?” And lo and behold, when I checked for monolayer formation three days post-seeding, I shouted, Holy Fucknoly! There it is!

    Honestly, it was probably the only time I literally shouted for joy during the entire 5 years of my Ph.D.

    Not really troubleshooting awesomeness, but in retrospect, troubleshooting is usually required because of some level of stupidity.

  7. Anne Says:

    I’m a technician that does plasmid construction. My first project took 6 months (although it should have only taken one). Every time I tried to gel purify fragments, ligate and transform them, I got nothing. Not even re-ligated vector. We tried everything. We made ligation buffer from scratch, tried 5 different types of competent cells, made new ampicillin plates, tried Ethidium Bromide and GelRed to see if the staining process was somehow to blame. After all this, one of my co-workers happened to look at the UV lamp we were using to visualize the gels and saw that it was short-wave and had been degrading the DNA. The first time we tried it with a long-wave, it worked perfectly, and we haven’t had the same problem since.

  8. Spiny Norman Says:

    What is a “library?”

  9. queenrandom Says:

    I spent 4 months of my life figuring out that to create an shRNA-expressing viral vector, one needs to digest not only sequentially (due to obvious buffer incompatibility), but it matters which enzyme is used first and which is used second, for no logical reason whatsofuckingever. Furthermore, trying to jump the process by designing your shRNAs with built-in half-sites is a fucking clusterfuck of fail. Also? Selective use of DNA phosphotase may work for normal directional cloning, however shRNA cloning, again for no logical reason whatsofuckingever, gives up the ghost and degrades itself during this process.

    It was super awesome to find out, after this ordeal, that half the shRNAs didn’t even work as published.

  10. Arlenna Says:

    Last fall, when we started our work, we thought we were having trouble making our peptides because our LC/MS data had these weird peak series that did not match up to our predicted masses. I stared and puzzled and was kept up at night about it for about a month and a half, I just had a feeling that there was something happening in the MS but I could not figure out what. Finally, one night when there was a snowstorm and I was stuck sleeping over in my office (because of how far away my house is from my campus), I stayed up until 11 pm playing around with spreadsheets. I constructed a table that considered a certain solvent-related adduct and all the possible species that could occur, and BINGO there jumped out the weird patterns we’d been seeing!! Now we know our peptides are fine, our solvent system and MS just does something funky during the ionization process that does not affect the actual material we have.

  11. Mad Hatter Says:

    One one of my first cloning projects, I was trying to clone a PCR product into pBlueScript KS. I always got colonies out of my transformations, but everytime I screened them by restriction digest, I would get this weird pattern–not the right pattern for plasmid plus my insert, but also not the right pattern for empty vector.

    I spent weeks remaking the PCR product, trying different restriction enzyme combinations, testing different vector:insert ratios, making new antibiotic plates, etc. None of that helped. Eventually I decided, screw it…I’m just gonna sequence a few of the clones.

    Lo and behold…what I thought was BlueScript KS was actually BlueScript SK, which has its multiple cloning site in reverse orientation relative to KS. The postdoc who gave me the plasmid was apparently mistaken about its identity and I was too inexperienced to know there were different versions of BlueScript. Since then, I have never trusted any reagent someone else gives me without verifying it first myself.

  12. Anonymous Says:

    This also took 6 months of my life

  13. I hang out over in the physical sciences, and I have a couple doozies. The most recent involved a set of scintillator panels for cosmic rays that had way too much noise on the line. I went through all my standard electronics checks, all my standard light leak checks. Then we broke an optical fiber that we were using for some other tests, and the only replacement we had was made of the same optical fiber that led from the panels to the pmts. Bam! Same noise on the line. Turns out it was the fibers that were light leaking. Because, you know, the most useful thing in the world is an optical fiber that isn’t light tight!

  14. Patchi Says:

    After success with 3′ and 5′ RACE, I set about to assemble the complete coding sequence for the gene I was trying to clone. The overlap PCR gave me a band the right size, which I digested and ligated into the expression vector. I got colonies, mini-prep DNA gave me digested bands of the right size… but when I reinoculated those clones into larger cultures the plasmid DNA no longer showed the insert. I thought I picked the wrong clone, so I repeated the midi-prep. No insert. I picked two different positives from the mini-prep, same problem. I went back a step, picked new colonies… same shit: positive mini-prep, no insert for midi-prep. I started over from the PCR, but still couldn’t clone it.

    I figured it must be toxic and tried the TOPO kit. Worked fine, so I subcloned it into the expression vector after I verified the sequence. Still couldn’t get the midi-prep to work. I got DNA, so I knew it wasn’t the kit. One day I joked that the bug was kicking my gene out of the vector… but somehow it liked the TOPO vector. Then I realized that the cells were different! I was using our homemade JM109 cells for the expression cloning, but TOP-10 with the TOPO kit! Once I switched competent cells the gene stayed in the vector. Weird shit…

  15. TomJoe Says:

    When I first got into graduate school I was tasked with picking up a project from a graduate student who was leaving. He had made mutants throughout the genes of a polycistronic operon (siderophore production) and we were going to do animal studies. I needed to ensure that the mutant we were going to use was truly devoid of siderophore production. So, I grew up the mutant which supposedly contained an insertion in the first gene, and checked it for siderophore production using the Csaky assay (checks for hydroxamates). It came back a weak positive.

    Go to the boss and show him the results. He doesn’t believe me. I go back and repeat the results, each time getting the same result. Boss doesn’t believe me. Heck, I’m a new graduate student and can’t possibly be doing it right, and the guy I picked up from was now a Ph.D. so who f**ked up? Couldn’t be him.

    So … I start doing Southern Blots. This was before massive DNA sequencing efforts so while we had the operon itself, that was about it. So between restriction digestion and Southerns, I was able to determine that the insert was actually in the SECOND gene of the operon. So the first gene was making a product which resulted in the weak positive reaction. I went ahead and made a new mutant and Viola! Csaky assay was negative. We then proceeded with the animal experiments.

    He shoots! He scores!

  16.’s done it again! Great read!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: