Tuesday, April 19, 2016
This Customer is Unsatisfied
Before actually starting to draft the Mod2 research report, I felt incredibly rich. Rich in time, rich in data (the whole class’s data!), rich in figures, captions, and results from previous homeworks, rich in ideas of all the different ways the data could be used. I felt like I had a pretty clear understanding of the NHEJ assay and the flow cytometry data, since I spent much of the past year making a similarly designed fluorescent reporter and running flow cytometry experiments in my UROP lab. Plus, preparing a journal club presentation, and then listening to all the other journal club presentations on NHEJ related subjects, made me feel like I had a solid platform of context in terms of other NHEJ research going on and areas of significant. All this lulled me into a (very) false sense of security. All the data was sitting in front of me, I had some directions in mind, it wouldn’t be that hard to put it all together.
I was very wrong.
It was very hard.
I took a closer look at all the numbers, both from the NHEJ assay and the drug dosage assay, and realized, “Wait, these numbers… these numbers are everywhere, these numbers cannot help me?” Numbers were conflicting each other, numbers were jumping around in strange ways, numbers were making very sad looking graphs, with error bars skyrocketing and increasing the whole y-axis and making my actual data look very small and insignificant.
Part of the frustration was the feeling that most of this funky data was not the result of biology, and maybe not representing what was actually going on in the cell, but the result of small cell counts or not enough replicates, or non-optimal experimental conditions. Were some of the numbers for the flow data so abnormally large because too many cells died from the lipofectamine? Or because the transfection efficiency was low? Maybe adjusting the plasmid:lipofectamine ratio would have given better transfection efficiency, or we could have incubated the cells for longer than one day after the transfection, to allow more time for cell machinery to kick in and start pumping out the fluorescent proteins. All these thoughts made it very difficult for me to accept the numbers I had, and work with them to create a story that I felt was scientifically sound and compelling. Yes, we know that sometimes we get unexpected results and must work with whatever results we have. But in this case, I couldn’t shake the feeling that these “unexpected results” were also unreliable. If these experiments were repeated, would they yield the same results again? Maybe they would, maybe they wouldn’t, but either way, this was the data I had for this report.
In the end, a large amount of the time spent on the Module 2 report was organizing and reorganizing the data I had in an effort to try to find a path towards a conclusion that was actually supported by the experiments and data present. I switched directions about four different times, and spent probably much too long simply agonizing over how to synthesize all the different and conflicting results into something that made sense.
I think the lesson learned here, is to look ahead at all the different experiments that will be done, and think about what information each experiment will give. Have a scientific question in mind that each experiment will contribute towards answering. This way, the question (and the answer) can evolve as experiments are performed and new information is found. Do not wait until the end, when all the data is collected and it is time to create a cohesive storyline, to really start thinking about what purpose this information serves.