Magazine: How Katie Pollard '95 Finds the Infinitesimal Genome Differences Between Humans and Chimps
When Katie Pollard ’95 was a postdoctoral researcher at UC Santa Cruz, she didn’t get much time with her boss. Even in 2004 David Haussler was already a macher in the world of biology. His lab had been a leader in the effort to sequence human DNA, and is the home of the UCSC Genome Browser, a sort of Google Maps for genomes. It was a big lab, and Haussler traveled a lot. Pollard didn’t--she had spent the last eight months trying to convince a powerful cluster of computers to find the infinitesimal genetic differences between human beings and chimpanzees.
It was a massive, punishing project. The fundamental unit of a genome is a gene, made of DNA, and the fundamental units of DNA are called bases—the A-T-G-C code that’s the blueprint of all life on Earth. The human and chimp genomes are about 3 billion bases long, but if you put the two side by side, just 15 million of those As, Ts, Gs and Cs would be different. That is, people and chimps are 99 percent identical, and finding the 1 percent that’s different is like trying to find a typo in an encyclopedia.
More Winter 2011 Pomona College Magazine Articles
- The Lives & Times of Jonathan Lethem
- The Medical Crusade of Dr. Emil Kakkis '82
- How Katie Pollard '95 Finds the Infinitesimal Genome Differences Between Humans and Chimps
- The Future of the Filibuster
- Kathleen Supove '73 Is Exploding the Staid Piano Recital
Download the entire winter 2011 Pomona College Magazine here [pdf] .
But Pollard had wrestled UCSC’s computers into submission. So when Haussler happened to be in the lab one afternoon as she’d just pulled some new results off the cluster, Pollard took advantage of the opportunity. She called him over to her desk and told him she thought she had something, a human sequence of just 118 bases that looked to vary a bit from the equivalent stretch in a chimp. Pollard pulled it up on her computer screen. “Well, why don’t you click on that first one?” Haussler said. The sequence was a gene--though it mapped not to a protein, as genes typically do, but to the alternate genetic material RNA. (Cells use DNA as a guide to make a chemically similar copy called RNA.) The gene was nearly identical in chimps, mice, rats and chickens, which meant it was really old (evolutionarily speaking), and probably important. But in humans, it was different. Pollard checked the database to find out what cells in the body used the gene, and it showed up in the human brain--the hippocampus, to be exact. If you were looking for differences between people and chimps, that’s certainly one of the places you’d hope to find them.
Haussler looked at Pollard and smiled. “Awesome,” he said.
They named the sequence HAR1, for Human Accelerated Region. It turned out to be the first of dozens of stretches of DNA unique to humans, and Pollard remains one of the best researchers in the world at finding them. Her background--a mélange of math, anthropology, epidemiology, statistics and biology--had turned her into a machine for squeezing knowledge out of the raw information of a genome. Learning to extract and decode those sequences was a hard problem that occupied much of the 1990s for biology, and only led to the even harder problem of the last decade, the one Pollard is working on: figuring out what all that data means.
Pollard now runs her own lab at the Gladstone Institutes at UC San Francisco, where she works on understanding what parts of the genome are found only in people, and what those genes actually do. “The chance of finding something like HAR1 is so close to zero that it’s sort of amazing this stuff even happens,” says Pollard. “The only way I can reconcile it is, there’s a lot of genomes out there, and a lot of time. With a lot of genomes and a lot of time, some random things actually happen. Really extreme, crazy things.”
GENETICS IS NOT DESTINY, but Pollard’s family might make you think otherwise. Her father, Tom Pollard ’64, is a Pomona alumnus, a cell biologist and the dean of graduate studies at Yale. Her uncles, Tom’s brothers Jim ’77 (a rocket scientist) and Dave ’65 (a geologist), also went to Pomona; only brother Steve (a mathematician) did not. “Every night at dinner was a science class,” Pollard says.
She had a knack for math in high school, and when she got to Pomona she found herself in high-level calculus her first year. She hadn’t really planned to pursue the subject, but her professors saw her talent for abstract problem solving. And anyway, she says, “there were two women in the Math Department, really good professors, and I’d never met a woman mathematician until then,” Pollard says. “My teachers in high school were all dudes.” Plus, Pollard was in an all-girl punk band called Fox Force Five--she played guitar--and one of those professors played music, too. They even jammed together once in a while. “It made me realize that it wasn’t totally taboo to be a mathematician.”
But it wasn’t totally satisfying, either. Jim McKenna, who’s since left Pomona for Notre Dame, got Pollard into biological anthropology--looking especially at childcare differences among primates. “But at the time I saw almost no connection between the anthropology and the math. They were just both things I liked,” she says. After college, Pollard took a grand tour through science. She got a Watson fellowship, went to Europe, got interested in epidemiology and public health, and made her way to UC Berkeley for a Ph.D. in biostatistics.
This was the late 1990s, and the genomics revolution was beginning. Pollard did a summer internship at biotech company Chiron, where the big push was making DNA array chips, hybrids of silicon and biochemistry that could tell you which genes in a cell were turned on--making protein--or turned off. And when the first rough draft of the human genome was released in 2000, the whole San Francisco Bay Area seemed to light up, drunk on the dot-com bubble and the apparent potential of biotech.
The hangover was brutal. “We thought that when we got the human genome sequence, everything would be kind of done, that it would be a simple matter of glancing into that genome and seeing the answers to all these questions we couldn’t answer before,” Pollard says. “It was a little bit conceited. We got the whole sequence, and we could see when certain parts were being made into RNA or not, but we still had no idea how the system operated.”
She got a job in Haussler’s lab, working on comparing newly sequenced genomes to each other. The idea is that if you find genes that are very similar across very different species, those genes are very old and therefore probably very important. Biologists say these genes are “highly conserved.” But really less than 2 percent of a genome codes for actual proteins. About half seems to be leftover from viruses that have invaded the mammalian genome over millions of years--no one’s really sure what all that does. And the rest, the so-called non-coding sequences, seem to control how much protein gets made from the genes at what point during an organism’s life and in what part of the body. And those sequences might also be more or less conserved across different species. Pollard’s job was to come up with computer programs that could automate the work of finding them. “I was excited to have her, because I knew she was brilliant,” says Haussler. “She immediately latched onto molecular biology and started to think of herself not as a statistician helping other scientists, but as a scientist in her own right.”
“There were a bunch of genomes being sequenced--mouse, rat, dog, chimp--and they basically said, ‘Katie, you’re new. Everyone else is pretty busy. Why don’t you get on the chimp?’” Pollard says. “And I thought, oh my God, this is it. It was a synthesizing experience for me.” Her amble through the sciences now made sense—biology at home, math in high school, math and primate anthropology in college, statistics in grad school… you could use math to study evolution. If you wanted to figure out if some sequence of DNA had fewer substitutions than chance alone could explain, you had to use stats, and a massive computer cluster, but you had to understand the biology, too. A field that Pollard didn’t even know she was studying finally had a name: comparative genomics.
POLLARD WENT ON to help identify 200 human accelerated regions of the genome--all non-coding, which means they don’t make proteins (though they might control how much, or how often, genes do make proteins). They all have obscure names, but their functions make intuitive sense. A year of “wet biology,” working with HAR1 in the lab, suggested that it was involved with the development of the cerebral cortex.
HAR2 has something to do with the shape of the wrist and thumb. Researchers in other labs focused on fast-evolving genes: FOXP2 makes the mouth able to form the complicated shapes necessary for words, LCT lets adults digest lactose so they can drink milk, and AMY1 makes an enzyme that digests starch and allows for a broader diet. One by one, these chunks of genome begin to paint a picture of what it means to be human--of what genes and non-coding sequences changed, fast, to divide us from chimpanzees 6 million years ago.
Of course, human uniqueness doesn’t privilege Homo sapiens, particularly. Look in the right places and you’ll find “chimpanzee accelerated regions” or even “beetle accelerated regions.” “We have these uniquely human bits, but you can do the exact same analysis and flip the role of human and chimp,” Pollard says. “You can find the parts where humans look like all the other animals and the chimp is unique.” In fact, she adds, for about half the human accelerated regions there are humans who have the chimp versions in their DNA (though you probably wouldn’t know from looking at them).
Finding the functions and meanings of all these pieces of the genome is called annotation--putting nametags on all the parts, understanding what each of them does, and describing how all of them interrelate. And genes are only the beginning of the story. “The differences among species in general, and especially between humans and non-human primates, are non-coding. It’s not genes but regulatory regions,” says Nadav Ahituv of UCSF’s Institute for Human Genetics. Those differences, he says, come from recombining existing elements of a genome, not waiting for mutations to generate entirely new ingredients. “If you change a gene, it could be a pretty drastic change. But if you change the timing or the amount of it, that’s less drastic. You can change small things.”
That makes the differences among species less a matter of who has which genes and more a game of recombination. “A good analogy is language,” says Jim Noonan, a geneticist at Yale who, like Pollard, works on HARs. “You put different combinations of words together and you get different meanings. But the words themselves are the same.”
The main scientific challenge now is to translate this new understanding of genes and evolution into an understanding of people’s physiology and health—to go from genotype (the genes you have) to phenotype (what those genes do). “We have really cool methods. We have data. We have fast computers. And we can find the uniquely human parts of our DNA,” says Pollard. “But that still doesn’t tell us necessarily what has to do with human evolution, or with things that we know are unique about humans, like speech, writing, music, programming computers or human diseases.” For example, both chimpanzees and humans can be infected with the human immunodeficiency virus, but only humans get AIDS. Blood lipid levels that would turn a human into a walking heart attack don’t faze chimps at all. Clearly we humans have had to take on certain weaknesses as the price for the ability to digest rice or use Twitter.
Luckily, the tools for studying evolution among populations also work for studying individual variation—which researchers like Pollard hope will lead to figuring out how the genome influences human health. The idea is to find people with particular problems, like serious obesity, and then figure out the tiny differences between their genes and someone skinny. It’s the same statistical problem as comparing a human genome to those of other species. “I’ve realized in the last year or two that it’s a big step from a screenshot of a gene sequence to a phenotype, and phenotype’s where we want to get,” Pollard says.
That’s a big job, of course. Pollard’s lab, in UCSF’s brand new Mission Bay campus south of downtown San Francisco, shares thousands of square feet worth of tall black shelves and equipment with a bunch of other researchers, post-docs, grad students and so on. She’s in charge of 11 people--an eclectic group of physicists, statisticians, biologists and computer scientists. Between that and her recent engagement, Pollard has little time to DJ at dance clubs the way she used to, much less play guitar. But it’s the kind of flitting among disciplines that Pollard is used to. “Everyone’s got their special stuff,” she says. She’s talking about genes, but it’s easy to hear something more profound. “Mostly we’re all really similar. And then we’ve got a few things that make us special.”