When It Rains...

This week was something of a study in how many things could go wrong in my experiments again. The week started out all right, because I was just doing things that I’d already done. However, on Wednesday we began optimizing our experiment so we could get conclusive results about whether our protein was present in the sample. You wouldn’t think that optimization would cause things to go worse... and in most cases you’d be right. But in this case, that’s just what happened. We started optimizing, and things started going downhill.

Optimizing started out fine- we found a paper that had shown the presence of our target protein in a different section of the body, so we started using their conditions. Almost immediately though, things started to not work as expected. There’s an assay (basically a fancy name for a test) that we do to figure out the concentration of protein in a given sample. The way the assay works is that you add a set of buffers (specific chemicals for keeping the pH constant) to your sample and then run it through a machine that sees how much light is absorbed by your protein, which can be mathematically translated into how much protein is in the sample. When it works properly, the buffers change the color of the samples such that there is a constant color gradient. In our tests on Wednesday, not only was there not a constant gradient, but the set of samples that were supposed to be entirely clear (because none of the protein used to get a baseline reading was in those samples) were some of the darkest purple on the plate.

Unfortunately, this happened at the end of the day, on a day when my postdoc had gone home sick, so there was not really anyway to figure out what was going on. I’ve learned a lot so far, but not quite enough to figure out why my tests don’t work without significant research. As a result, corrections had to wait until the next day. My postdoc also had very little idea what was going on with the assay, so I just ended up redoing it, hoping that I had done something wrong so that it could be corrected by running the assay a second time. However, this hope was unfounded. As it turned out, running the test a second time gave me even more wonky results. In theory I should have gotten a straight line. What I got instead was a zigzag. Not exactly helpful when you’re trying to approximate the slope of the curve with a straight line. We did determine that it wasn’t my fault the test wasn’t running correctly, which was something of a relief (because I hate messing things up). It was still mildly frustrating that no answers were available though.

Since we couldn’t get any useful information out of the test, and we still weren’t (and honestly, still aren’t at this point) sure what was going wrong, we decided to move on. There’s another, less exact, way of measuring protein concentrations that involves running differing dilutions of protein out on a gel and matching the band size to a known concentration. We diluted the protein appropriately and started running the test. Unfortunately, we didn’t see any of our target protein in the resulting blot (basically a mirror image of the gel that can be read on a specialized scanner). This wasn’t totally unexpected, since we haven’t seen any of the protein in previous tests, but it wasn’t what we were hoping for. However, that portion of the experiment went all right, so things going wrong hadn’t quite started yet.

Things started going really wrong when I started running tests to see if there was any protein, of any kind, present in our sample. There are types of protein that are expressed (present) in all cells, regardless of where the cells are coming from in an organism. In order to see if we had any proteins, we added antibodies (cells that bind to a specific protein) that would bind to one of these ubiquitous proteins in order to see what was present. So far, so good. (At least theoretically. At this point in the process of this test, you’re pretty much flying blind about whether or not anything has happened... there’s no way to see what’s going on until later.)

Things really went wrong when I added the secondary antibody (the one that would bind to the first antibody, and has a fluorescent tag that will glow in the special scanner and allow us to see that there’s something there. My postdoc has been trying to write her paper, and I’ve been trying to take initiative, since I theoretically know how to do things in lab now. It’s been five weeks after all. Usually, my initiative-taking works out in the end. I might take a good deal longer to do things, but I get the result that we’re looking for. This time, not so much. When choosing the secondary antibody, I picked one that would bind to the primary antibody (which was correct) but I didn’t pick one that would glow. This meant that we saw a whole lot of nothing when we looked at our results, which was pretty much exactly the opposite of what should have happened.

Since there was now something bound to all of the antibodies we wanted to test, I couldn’t just add the correct antibody. So, today’s (Monday’s) project was starting the entire thing over again. It’s been ever so much fun. There were a couple of morals I took from the whole disaster though. First, ask questions if you’re unsure. The only stupid question is one that doesn’t get asked when it should have been. And second, the old adage about measuring twice to cut once is apt no matter what you’re doing. Double and triple checking the instructions and your ability to follow them will save a good deal of time in the long run. I suppose that since part of learning seems to be messing things up, I’ll have to get used to it... I’ve still got a lot to learn, so I can only imagine the mistakes I’ll make along the way. :P