Have you ever found yourself in a place where you asked yourself, ‘well how the fuck did I end up here’? Well, I did today, and god, this day was boring as fuck.
As part of an experiment on the effects of light pollution, which I conducted at the end of 2020, I was interested in how the expression of certain genes would be affected by my light treatments. To do so, I used qPCRs. Or well, I’m not a molecular biologist (at least not anymore), so one of my collaborators – who happened to be a molecular biologist – took that role upon them. Over the first half of this year, we extracted mRNA, and sent the samples over to our collaborator. Just before I went on parental leave, I received an all worked-out version of the data. I had a brief look, and then forgot about it for three months. Given that the ecological data from that experiment are very exciting, I was very keen on picking up the qPCR data again, and so I did that today.
I had forgotten what a horrible experience it is to work with qPCR data. I have worked with it in my studies, and for a paper I published in 2018. The steps involved are not that hard, but there are plenty, so it is easy to get lost. Because I got a worked-out version, I missed several of these steps, critical to my own understanding. As I couldn’t follow the all worked-out data, I insisted with the collaborator that I needed the raw data, so I could calculate everything myself. It took me many hours today, only to end up with exactly the same figures. They did it right. This is what I knew already, but I’m of the opinion that if you write about something, you should more or less understand the principle. That’s the main reason for going through the whole procedure from scratch. This helped me understand the process – again.
This may not even be the worst part. The worst part is interpretation. What the fuck does it all mean?
I made a few quick figures, and my plants expressed completely the opposite of what I expected them to… Given that the calculations involve several changes of sign, this almost makes me doubt my own thinking process. Did I mess it up somewhere along the way? Nah, I’m pretty sure the calculations are solid. Especially since the experienced collaborator came to exactly the same conclusion. Nevertheless it’s another rabbit hole to dive into… That’ll be for Friday.
Another pretty terrible aspect about gene expression assays is the variation. Even though we did our very best to reduce the contribution of single outlier individuals, by pooling material from three replicated plants within each experimental treatment, the variation looked considerable. Each qPCR was performed three times on a sample – the technical replicates – which were pretty stable and similar. Regardless of all the efforts to reduce confounding variation, it seems to me that the variation between the true replicates is all over the place.
But what do I know, I haven’t analyzed the data statistically. I might as well be wrong… We’ll see soon enough.
That’ll be it for today, me expressing my feelings about plants expressing their genes. Let’s hope I can – again – steer clear of it in the coming years.