What is the future for the analytical laboratory? Time to extrapolate
21 Sep 2015 by bpbbjpttc.d bpbbjpttc.d
Are we to trust the extrapolators of the world? And if so, what does this mean for the analytical laboratory of the future? Russ Swan has some answers…or rather, some more questions.
Are we to trust the extrapolators of the world? And if so, what does this mean for the analytical laboratory of the future? Russ Swan has some answers…or rather, some more questions.
There are laws, and there are laws. The immutable laws of nature, such as light speed being the absolute maximum velocity, and every action having an equal and opposite reaction, seem to be unbreakable by anybody. Anybody except a science fiction hero.
The laws of the land, on the other hand, are breakable by everybody who neglects a speed limit or prefers a non-approved recreational lubricant.
In between these extremes are manifold 'laws' of science – the ones that seem to be true, so far at least, but which anybody with any sense would take with a pinch of salt. A number of these soft laws are based on the extrapolation of observed trends rather than any fundamental understanding of underlying mechanisms.
Perhaps the best example is Moore's Law, the one about the number of transistors that can be crammed onto a computer chip doubling every couple of years or so. Its predictions have held true for decades now, driven by the market dynamics of the electronics industry and – perhaps – by the helping hand of self-fulfilling prophesy.
The logic goes like this: Every executive and every engineer at every chip manufacturer is aware of the law, and knows that every one of their competitors is too. They all recognise that they have to meet the future targets set by the law, or they will appear to be falling behind, and so they make damned sure that they do actually double the power of each new generation of processor.
And what started out as an interesting observation about past improvements in technological capability has become a roadmap for the future development of advanced industry. What would have happened if Gordon Moore had never published his thoughts, or if he nominated the doubling time as five years – or six months?
Extrapolation, as we were all taught at school, is a dangerous thing. The economist Karl Kapp noted in 1972 that if a computer had existed one hundred years earlier, it would have predicted that the observed growth in the number of horse-drawn carriages would result in the entire planet being submerged under ten feet of manure by then.
I'm a big fan of Hooke's Law, and would even go so far as to say that it is my favourite of all the laws of science. It describes the simple straight-line relationship between stress and strain, showing how most solid materials stretch or shrink a predictable amount under elastic deformation. Up to a point.
Once a certain limit is reached, the straight-line graph goes all over the place, wobbling wildly until it ultimately ends in the destruction of the sample. It's an elegant example of the power of scientific observation, and of the dangers of hubris when making predictions based on extrapolation.
The Hooke's Law Analogy really ought to be applied much more widely, and not just in science. I can think of a few politicians and economists that might benefit from an understanding of the concept of the elastic limit.
Having established that extrapolation is a thoroughly bad idea, I can't help wondering where current trends in laboratory instrumentation are taking us.
Analytical instruments are becoming progressively smaller and lighter, cheaper (a little!), with lower detection limits and greater resolution. In this regard they are following much the same trajectory as many other examples of advanced technology such as computers and phones: smaller, lighter, and more capable.
Instrument manufacturers cannot claim anything like the Moore's Law rate of performance doubling, although our industry can point to spectacular examples of exponential improvement in certain key sectors (like genomics, natch). We're also seeing the first widespread applications of some new and potentially amazing technologies like microfluidics, which will certainly accelerate some of these trends. Increasing automation is taking away much of the laborious preparation of samples and monitoring of instrument runs that keeps many lab technicians in employment, and the simplified presentation of output makes the interpretation of results something that is no longer restricted to the most qualified.
In other words, it doesn't take much crystal ball gazing to figure out that the analytical laboratory of the future will be unrecognisable from that today. Point-of-care diagnostics are already common and will become universal, while the devices we use will shrink to the size of a smartphone or smaller, with a point-and-squirt operation that can be understood by practically anybody.
There will be little need for laboratories made of bricks and mortar, or for armies of white-coated technicians. The multifunctional instrument of the future will make Star Trek's tricorder look quaint.
In fact, all we need is a few more doublings of performance and the standard greeting between people of our ilk will become “is that a laboratory in your pocket, or are you just pleased to see me?”