This year's Nobel Prize in Chemistry goes to three scientists who developed computer programs that have become as important to chemists as test tubes.
These programs accurately simulate how large, complex molecules behave. The work is central to drug discovery, materials science and much more.
Chemistry is all about the interactions of atoms. But these building blocks of all matter are far too small to see, even with the most powerful microscope.
So, over time, chemists have come up with two different ways of visualizing atoms and the molecules they form. And they have written computer programs to simulate each approach.
Simpler, classical models “treat the bonds between atoms as springs,” said Middle Tennessee State University chemistry professor Preston MacDougall. "Some springs are stiff, some springs are floppy. Some things twist easily, some things are harder to twist.”
Those models are fine for looking at how the shapes of molecules change in different conditions, for example, when they are hot or cold.
But they don’t tell you much about what happens when bonds between atoms break and reform, as happens in the enzymes that do the work in our bodies.
For that, you need much more accurate and complicated models that use quantum physics.
But using quantum models for all the bonds in a big molecule like an enzyme would take an enormous amount of computer power.
“In modeling, as in many aspects of life, there’s a certain ‘you get what you pay for’ aspect,” says University of Minnesota chemist Chris Cramer.
Beginning in the 1970s, Harvard chemist Martin Karplus, Stanford’s Michael Levitt and Arieh Warshel at the University of Southern California developed computer models that successfully combined the two.
“They figured out a way to have a connection between the two that would let you have really fine-grained focus on an interesting piece of the big system, while not spending as much (computing power) to include the larger system,” Cramer said.
For example, he said, computer programs today use quantum models to study how a drug will react with the small part of an enzyme that performs chemical reactions. But the programs use simpler classical models to understand how the rest of the enzyme interacts with its surroundings.
These models have proven extremely valuable across the field of chemistry. For instance, scientists designing solar panels use them.
“Certain atoms are absorbing light and undergoing transitions which you must use quantum theory to model,” said Preston MacDougall. “But then, you also want to describe the plastic material that it’s embedded in, so that their bending properties are modeled properly, their thermal expansion, their mechanical properties are also modeled correctly.”
These models are so good that they accurately predict what happens in real life.
And according to the Nobel Prize committee, “Today the computer is just as important a tool for chemists as the test tube.”