Wednesday, December 2, 2009

Einstein meets Newton, or Diabetes and the Theory of Relativity

Isaac Newton was one of the founders of modern science. He helped promulgate the idea of formal Laws of Nature, that Nature was law-like in that everywhere and every time the same fixed principles were at work, like the totally deterministic Law of Universal Gravitation. Because the Law worked everywhere and was deterministic, it worked at the smallest possible level, such as in time or space. That's why Newton could formulate natural law in terms of the calculus or, rather, that's why Newton developed what we call calculus. And this worked in part because space and time were absolutes. In that sense, causation was absolute as well.

Charles Darwin was clearly a product of the Newtonian age. He repeatedly wrote about natural selection as if it were at Law of Nature -- one Nature, indivisible, with Liberty and Justice for all. Justice here, would be the realization of your inherent fitness.

But life is not necessarily like that. Einstein spoiled Newton's party by showing, so to speak, that things were not absolute but only had meaning in terms relative to each other.

In evolution, we usually frame natural selection in relative terms -- this variant at a gene does better than that one does in this particular population and time. We could also view evolution in more absolute terms: this species or population, as a whole, does better today than it did yesterday. Usually this would refer to population expansion which would come at the expense of some other species, so would be relative in that sense.

In genetics, we tend to want to be Newtonian/Darwinians. We want genes to cause things in a rigorously predictable, law-like way. Thus we think of genetic variants as having inherent fitness value. And the important area of biomedical genetics is GWASh in such deterministic thinking.

But life's not like that! There is, for starters, a substantial number of 'known' disastrous disease-associated variants in humans that are normal in other vertebrates (one estimate is that 10% of our 'disease' mutations are like that). In fact, it's true even within our own species: in every individual whose whole genome sequence has been published to date, there are many 'disease' alleles, though the person is unaffected as of his current age.

And then there's the environment, which geneticists pray to Newton will go away! Genes have their effect in context, a point so obvious that it's easy to ignore if it's bad for business.

An area of Finland had the world's highest recorded risk of heart disease. But over a 20 year period, change of diet was the major factor in reducing heart disease mortality in Finland by something like 65% (that's no typo!).

And a new paper in The Lancet on diabetes, comparing outcomes in progression to diabetes in at-risk adults reports that "intensive lifestyle intervention" was significantly more effective in preventing diabetes than medication or placebo.

We've talked of Newton and Einstein, and how causation is relative. But even Einstein hungered for Newtonian lawlike deterministic Nature. And along came quantum mechanics.....

No comments: