How Do Living Things Work?
Most living things are made of cells. These cells contain millions of tiny machines called proteins that are so small they can’t even be seen with an optical microscope. Proteins keep themselves busy with all sorts of tasks, from performing chemical reactions to receiving signals from outside the cell to helping other proteins communicate with each other.
Most of these proteins are very helpful, but every once in a while there’s a protein that does something unpleasant, like a virus protein that makes people sick or a cancer protein that tells human cells to keep growing into a tumor. Wouldn’t it be great if we could stop these proteins from working? We can!
Throwing a Wrench in a Protein Machine
It turns out that most proteins have small pockets that are sticky. Specific “molecules” (e.g., drugs) made of only a few dozen atoms bind to these pockets if they have just the right shape. It’s kind of like the way a key fits into a lock. Not just any key will fit.
It’s very difficult to find a molecule that sticks to a disease-causing protein. Often a super computer thousands of times more powerful that the one at your house has to be used. Fortunately, when we do find just the right sticky molecule, it often stops the protein from working. If the protein being targeted causes a human disease, a tiny little molecule can improve human health by altering the undesirable protein’s function!
A Key for a Lock that’s Always Changing It’s Shape?
Part of the reason it’s so hard to find one of these sticky molecules is because the protein pocket isn’t sitting still. It’s constantly changing it’s shape. Since it’s difficult (sometimes impossible) to see how proteins move, we have to perform computer simulations to predict their movement. Once we understand how the protein pocket jiggles, we can better design molecules that will stick to it.
It turns out that it’s EXTREMELY difficult to predict how proteins move. It takes tremendous computer power far beyond what most people have encountered to perform these kinds of calculations.
By studying protein jiggling and molecule sticking, I’ve discovered molecules that might one day be turned into treatments for diseases like cancer, malaria, African sleeping sickness, and Chagas’ disease. It’s great fun using computers to try to help people!
A computer simulation of a disease-causing protein
Wait… Even if You Know How a Protein Jiggles, How do You Know if Your Molecule will Stick to It?
The truth is, we don’t know. Computers aren’t yet smart enough to say with certainty whether or not a molecule will stick. However, computers are pretty good at guessing which molecules are sticky. In the end, we have to perform a real-life experiment to test the computer’s guess. Because the computer is good at guessing, though, it’s recommendations ultimately make it so fewer molecules need to be tested before a good one is found. That can save tons of time and millions of dollars.
An important part of my research is teaching computers to be better guessers. I use artificial intelligence (a “neural network”) to create a miniature version of a brain inside the computer. I then teach the neural network to identify sticky molecules by showing it lots of pictures of molecules sitting in protein pockets and telling it which ones are the stickiest. As the neural network sees more and more pictures, it gets better and better at picking the good molecules. It’s kind of like training a dog, if the trained dog could then be used to save millions of dollars and cure tons of deadly diseases. That would be an awesome dog.
So now you know all about my research! I hope it’s gotten you as excited about science as I am. Being a scientist really is wonderful. Maybe it’s the right career for you too!
This article is obviously geared towards a general audience. Scientific colleagues might be interested in a more technical overview of my research. I recommend perusing my peer-reviewed publications to learn more about my work.