George Church is a professor of genetics at Harvard Medical School and a professor of health sciences and technology at Harvard and MIT. In 1984, Church, along with Walter Gilbert, developed the first direct genomic sequencing method and helped initiate the Human Genome Project. Church is responsible for inventing the concepts of molecular multiplexing and tags, homologous recombination methods, and DNA array synthesizers. Church initiated the Personal Genome Project in 2005 as well as research into synthetic biology. He is director of the U.S. Department of Energy Center on Bioenergy at Harvard and MIT and director of the National Institutes of Health Center of Excellence in Genomic Science at Harvard, MIT and Washington University. He is a senior editor for Nature EMBO Molecular Systems Biology.
Question: When did you first incorporate computers into your work?
George Church: Well the first connection was really between biological polymers – the things that make up our body – DNA and proteins. They’re very much like a computer program or a computer set of data. Zeros and ones turn into As, Cs, Gs, and Ts. And that’s . . . once you see it, it’s such an obvious mapping that seems very profound. And it’s as if somebody’s been programming all this for a long time and has done a really nice job of it, and you want to look at it and appreciate it the way you would appreciate any work of art or design. So yes, I think . . . Then it turns into a three dimensional structure. So you have this simple binary or simple digital stream of As, Cs, Gs, and Ts, and it turns into this folded molecule. And that’s what we were doing in crystallography was looking at the three dimensional structure. And that was an interesting metamorphosis.
Well you can think of crystallography as kind of a way of taking pictures of very, very, small things. It’s smaller than you get with a microscope; smaller than you get with an electron microscope typically, at least back then; but it’s more complicated because instead of just taking a picture, you make this diffraction pattern. And then you have to use a computer to grope it back to something that’s more like a three dimensional image that humans think about and that is actually the reality of the . . . And it gives you atomic positions of every piece of the molecule. And so we were really pushing the envelope back there in the mid ‘70s by studying what’s now called “macromolecules” or the proteins in nucleic acids of life, rather than what had been so that you could do x-ray diffraction on salt crystals, like table salt. This was much, much larger molecules that we were dealing with.
When we – and by “we”, a large community – saw the first examples of folded nucleic acids, it really started a series of events that . . . It’s like watching a _______, the first DNA folding. We saw the folding of a _______, and it said that the code for going from the simple digital molecular description – its name, basically . . . its sentence structure – to this fold of molecule, it’s something that you could do on the computer. You could either solve it theoretically or you could solve it by x-ray diffraction. And I think that knowing the parts, the shape of the parts that make up our body . . . It would be like trying to work on an engine without not being able to know what it looks like or what it feels like. In other words, even a blind person knows the shape of the parts of a car, and we didn’t know the shape of anything that we are made out of. So that’s now changed radically. We know the shape of most of our parts.
Topic: How did new technological developments change the way you worked?
So from that point, it seemed reasonable that the kind of computing and automation that was used in crystallography in the mid-70s could be applied to the rest of biology. And so I restarted graduate school with Wally Gilbert, who a few years later got the Nobel Prize in chemistry for doing the first kind of really reasonable sequencing method that could be scaled for DNA. And we started dreaming about ways that you could automate the process of collecting the sequence data, meaning the order of As, Cs, Gs and Ts for DNA, and quantitating how that plays out in terms of the amounts of proteins and other products of the DNA in your body and in various cells. And so that automation of analyzing the DNA and the _____ and the proteins, then transition more recently to synthesis. So automating the process of synthesis where you would take what you learned analytically, and turn it into sort of nano scale sculpturing . . . making machines where now you know the parts; now you can put parts together in new, and interesting, and useful configurations. So I think that what I see is sort of the spark that happened in the mid-70s played out into automation and computer analysis, and kind of this very intellectual task of figuring how it all folds up, how it fits together, and how to make it useful, making the biotechnological companies and a shared academic experience.