I’ve just walked out of a wonderful meeting that has kind of left me on a “science high”. The Raj Lab (my new scientific home in Philadelphia), Thomas Gregor’s lab from Princeton and Dan Larson’s lab from the NIH had a get-together, talking science, methods, data. It was really great, with loads of lively discussion. But I don’t want to discuss any of the actual science here. Instead, I’d like to share some ideas about scientific terminology, how a given word might have different “baggage” attached to it depending on your background/training, and whether fuzzy definitions may actually be useful in biology.
I started thinking about this, because we were looking at a lot of microscopy pictures, like the one above. These pictures generally showed fixed cells where transcribing mRNA was labelled, and could be seen as little dots. A generally accepted view of transcription is that it happens in “bursts”, and certain parameters of these dots (e.g. their number or intensity) can be considered a proxy for different parameters of the transcription process. Therefore, when people were describing these pictures, they often didn’t refer to the spots, but the transcriptional parameter. For example, they’d say “we see bigger burst size”, instead of saying “we see a brighter spot”. So, at one point someone commented that this was not good scientific practice, because by doing so, we were not giving listeners the “raw data”, but instead we were giving them pre-interpreted results, evoking ideas they might have attached to that term and thus biasing them.
The following discussion revealed that talking about bursting did indeed evoke different interpretations in people with different backgrounds. People with a more physics-y background considered bursting to represent a very specific model of transcription, whereas people who had more of a biology background were considering bursts as a way to describe the cell-to-cell heterogeneity in a cell population.
And I’ve noticed that people with different backgrounds frequently attach different meanings to the same scientific terminology elsewhere, too. Here some recent encounters I’ve had:
– stochasticity, randomness & (un)predictability: some scientists will point out that there are precise mathematical definitions attached to stochasticity and randomness, and these are not at all identical with the predictability of something. Moreover, we often apply any and all of these terms very loosely when describing biological systems, without testing if the mathematical criteria are met.
– mechanistic insight versus random events: at a recent meeting about the mechanisms regulating transcription, there was a lot of talk about the role of stochastic events. Some people in the audience were not scientists, and seemed rather confused by this discussion: for them, a “mechanistic understanding” meant a deterministic model, where it was inconceivable to have any kind of stochastic parameter.
– the robustness of a biological process: a developmental biologist might use this term to contrast developmental outcome (eg developing a foot) with variability at the transcriptional level. If the former was invariant, and the latter highly variable, one might say developing a foot is robust to changes in transcription. A molecular biologist working solely on transcriptional data, however, might talk about a “robust two-fold gene expression change”. Here, they usually mean that the two-fold change can be reliably measured.
So, is all this loose terminology bad science? You might say that ideally, whenever we discuss a scientific question we should precisely define the terms we are using. And maybe you are right. But I’d share a thought and an example that I heard from a professor at a meeting called “Chance at the Heart of the Cell”: Biology is a complicated process, and often when we start a project we don’t really know what we will find. So, maybe having fuzzy definitions is important in biology, so we can think out of the box a little? The professor gave the example of the gene: although almost every biologist will claim they know what a gene is, it is very difficult to exactly define it. Yet, precisely the absence of such a definition may have helped us to unravel the path from DNA to RNA to protein, or to learn about alternative splicing, pseudogenes and genomic duplications, or to discover compensatory mutations.
I’m not sure if he’s right, but I have learned that it is often helpful not be too precise with biological systems. Sometimes you can do the perfect experiment that should work on paper, and yet it doesn’t work in real life, and the only explanation you have is that “biology is complicated”. So maybe there is some higher level organization in biological systems, which is more than the sum of the parts, and requires some – not so precise – perspective?
I can very precisely tell you that… I don’t know.