What is a molecule?

I am a computational biologist. I compute molecules. I sit in the darkness of the computer lab and sculpt molecules by wielding the sharp scalpel of my mouse. Inside the computer I manipulate the coordinates of a molecule (currently the PDZ domain) and surrounding bath of water molecules. I apply force-fields, mathematical descriptions of the interactions between atoms, to my molecule, and hopefully, I coax the molecule into doing something remarkable: a flip, a rotation, a clamping of the active site - an action that might explain a chemical reaction in precise atomic detail.

This kind of precision is beguiling, and it engenders a kind of arrogance. That's why I often have to pinch myself hard to remember that, even though I might know the behaviour of these molecules down to the sub-Angstrom (0.00000001 m) level, I barely know anything about the molecule.

Chemicals start off as something in the ground, or in the ground up substances of animals. I rarely know whether the molecules that I study inside my computer comes from an animal, a plant or a bacterium. And what if I did? I still have to know how to squeeze that particular chemical out of the carcass of, say, a dead cow. This requires expertise in the manifold arts of physical chemistry, whereby you separate, from the rough and guts of a ground-up cow, the millions of different chemicals that make up the once living cow. Understanding a chemical requires not just in knowing how to find it in nature, but also how to purify it till it's purer than the driven snow. But to truly understand a chemical, you have to know what it does, how it reacts with other chemicals, and under what conditions.

In biochemistry, the problem of knowing what a chemical has taken a rather strange turn. We actually possess a rich source of important biological chemicals, but no way of knowing what these chemicals actually do. This source is the human genome, which exists as a publicly available database. As we know quite a lot about the grammar and syntax of DNA, a computer scientist can trawl the database for sequences of DNA that code for a completely novel biological molecules. It's then a simple matter of sending in an order for a biotech company to make the molecule from scratch.

But how to figure out what the molecule does in our bodies? We don't believe in mysterious life forces any more. We believe that all of life's processes, from digestion to respiration to the way old people lose their memories must rest on some kind of chemical process. Every chemical extracted from living things could potentially have an important function in the living process. But given a molecule picked out at random from the human genome, we have no way of knowing where the molecule should be found in the body, what biological processes that it takes part in, and what other processes it depends on. The human genome remains silent on such issues.

The precise knowledge of the 3-dimensional coordinates of a molecule is probably the last thing that scientists get to know about a chemical. 3-dimensional coordinates are finicky things, and knowledge of them normally comes at the end of a long investigative process. By then, much of the chemical properties of the chemical are already known.