Life of a Lab Rat

Heather is whinging about her spectrophotometer. The 'spec' is a machine that shines light of a defined wavelength through a sample and measures how much is absorbed. It returns a quantity known as 'absorbance' or (incorrectly) 'optical density' ('OD'). This is useful to us because the amount of light absorbed is directly proportional to how much stuff there is. Furthermore, different stuff absorbs different wavelengths of light.

So, to pick a completely non-random example, you could make some DNA, put some in a cuvette (the sample vessel, of a defined size) and analyse it in the spec by shining light of wavelengths 260 nanometres (nm) and 280 nm (these are far ultra violet wavelengths) through it. You would find out how much DNA you have, because there is a simple formula that relates absorbance at 260 nm to concentration, and how much contaminating protein there is, because (most) protein absorbs light with a wavelength of 280 nm.

Thus you can therefore measure protein concentration, or follow a colorimetric reaction by observing at the appropriate wavelength, and so on. A spectrophotometer is a very useful instrument, very simple in principle, these days very easy to use, and a source of great frustration to many.

The problem with it being seen as a 'random number generator' is one I have come across before. It is very easy to blame the spec for giving strange data, and to call in the rep and then get even more frustrated when they say that the spec is fine, thank you very much, look at my standards. But the problem is real, so if the instrument is fine (and let's face it, it is making a simple measurement; if there is no dirt on the sensor and the light source is good there is not a lot else to go wrong) then what is happening?

Heather makes the common mistake of assuming that the "only logical explanation" is that there is something wrong with the instrument. This is an extremely illogical deduction: Are we really suggesting that the machine knows when the rep is putting standards through it and when Heather is doing an experiment? I don't want to get too Qiagen on your arse, but this leaves operator error (unlikely; these things are dead easy to use), dirty cuvettes (likely, but the first thing any scientist worthy of the name would check) or sample preparation.

And it is this last that is the most likely cause of Heather's frustration. When we called in the Beckman rep to our lab in Oxford to check out the spec, he confided to me that the machine was fine, but that people did not prepare samples properly.

You see, DNA is long (for a molecule), sticky, and has bizarre properties. If you make a plasmid preparation, it will suspend in buffered water quickly enough, but it is by no means homogeneous. It takes a substantial amount of mixing to get the same number of molecules of DNA in each unit volume of solution. This is critical when it comes to measuring the absorbance. Dip the pipette into a freshly-suspended or prepared solution of DNA and you are highly unlikely to get the same number of molecules a second time. Because DNA absorbs ultra violet light really well, it is usual to dilute it before measuring the absorbance; and when you do this it takes a finite time for the concentrated DNA to disperse evenly throughout the solute. It takes even longer if you do not actually mix it. When the light beam is relatively narrow this becomes even more important.

I know scientists are always in a hurry, and I know that the usual method is to take the freshly-prepared DNA, dilute it, and measure the absorbance as soon as possible. And I know that this does not cut the mustard when it comes to getting precise measurements. My advice to Heather and Hal, as is often my advice in the lab, is to "vortex the crap out of " their samples.

That's a technical term, by the way.