N=4

Went to a talk yesterday and got stuck with the same question I have had several time after “changing fields” (not really but let’s say that it is more in the field of immunology and other mice related work where this issue comes up more frequently). The talk was sure enough based on preliminary data, and I am all for that if you state that, and somewhere in the middle of it the issue came up that all the data showed work based on n=4. How I realised that? Well, in the middle of interpreting a slide where the late days looked interested the speaker says “I am not sure I believe in this day since it is only based on one mouse since the other three in the group already had succumbed”.

This after I had heard “significant difference”… and I went into a loop of trying to remember my statistical training and how to get data significantly when comparing different groups of four with each other. I guess you can get significant difference, if one group is four and one is one… but if one is four and the other is three… I don’t know.

And somewhere there I remembered again the discussion I ended up having after another talk (biochemistry and structures) with one of my post doc colleagues, who is an immunologist, and the bashing of the entire field of structural biology and the interpretations of structures – especially when the protein in question doesn’t have a biological function yet. The statement from my colleagues was mainly that “it is not really something you can tell since you only look at this ‘structure’ rather than biological* assay” (*testing in cell culture or in vivo).

Personally, I like crystallography, and indeed the other talk yesterday I ended up thinking about was the one where they had investigated amino acid mutations and implications that made for virulence and pathogenesis but they did not even mention how the mutations changed the alpha helical structure, or the stability of the protein or something along those lines, i.e. biochemistry and chemical properties. It mainly makes me question how different we look at questions and interpret data beginning on where we stand – in which field as in immunology, structural biology or something like that. I am a little worried that the gap is larger than I previously thought.

Although, I guess I should remember my earlier training where I studied proteins and was trying to collaborate with both chemists and microbiologist as well as MDs (who in this case would study the biological implications in vivo). I guess the thing I wanted to say here was the following; significantly is a word I don’t like people using when it is clear that the statistical analysis is somewhat shaky, some fields come from so different angles that it is pivotal to stop for a second and reflect about what it means from other points of view and that not everything you are interested in, or think is obvious for that matter, is the same for other people who are trained in another discipline.

And I still linger with the thought that it would have been interesting, and so much easier to find a job, if I would have pursued that PhD in chemistry or maybe even made my training in a department called “biochemistry/chemistry/’something more clear’”.

Ah well, it’s the endless thoughts of “what ifs” and “how did they think there?”. Time to go to bed and not ponder n values and statistics too much, but I am happy that I managed to force myself not only through that [awful] statistics classes but also gain programming skills and bioinformatic training. Never have I seen such a need to pretend to have [actually have] skills as a chemist (or a civil engineer) or other “multitasking people who can be a good resource” and not ‘only a microbiologist who knows molecular biology’.

Definitely time for bed.