Interview with George Church published in SmartPlanet.
Harvard geneticist George Church has ambitious plans to sequence 100,000 genomes through his Personal Genome Project.
By Boonsri Dickinson
Long before he could grow his signature beard, geneticist George Church fantasized about sequencing the genomes of mankind.
Today, that dream is a reality.
Three years before anyone else thought to sequence genomes -- 1987, to be precise -- Church was in his Harvard University laboratory unraveling the DNA data code.
Hype is mounting for the 10-year anniversary of the announcement of the first draft of the human genome, officially this June.
But Church admits that he's not at all impressed -- despite $3 billion already invested, humanity is far from completely decoding the human genome.
Perhaps no one has seen genomics as up-and-close as Church, who became his own guinea pig in thePersonal Genome Project, or PGP. To date, the project counts more than 16,000 volunteers -- but only a select dozen has made their genetic and medical history public.
Eventually, 100,000 people will be sequenced through the project.
This week, Church is in Steamboat Springs, Colorado, where he'll be speaking at the FASEB Summer Research Conference on Genome Engineering. I spoke with him yesterday about why genetic engineering is about selective application -- and why the U.S. needs to defend its lead in genomic technology.
Now, how exactly are you going to program biology to do what you want?
There are two sides to my research: reading and writing biology.
The genome engineering is on the writing side. It's not so much about changing every part of a genome at once, but cost-effectively engineering many parts of genomes that matter -- for example, engineering them for resistance to all viruses.
On the other side, we are reading genomes for personal genomics. But there are many ways that reading and writing interact. We need to read the genomes of nature to get ideas then re-write these in novel combinations, and then re-read them to see that we made what we intended or see how lab evolution has made our constructs even better.
Did you learn anything from having your genome sequenced?
We are writing software for everyone to use. We have 32 public genomes already on which we are testing this software. We stare at the output from 32 genomes. Mine is probably among the most boring.
For example, PGP number 6 had the allele of hypertrophic cardiomyopathy. It's a condition, where one day you are shooting a basketball and then suddenly drop dead. It’s not part of standard medical care to look out for this. We recommended that PGP 6, get an endocardiogram.
It's a perfect example that can happen to you late in life. You can look at the many highly predictable and highly actionable gene variants like fire insurance. Everybody is at risk. You can't say that you aren't at risk until you look at your genome.
Can you tell me more about how you were among the first to start sequencing the human genome?
As a teenage researcher in 1974, I typed in all the DNA into a computer. At the time, almost no biologists used computers. I thought it would be nice to take these strings of [the letters of DNA] A, C, G , T, and fold them up and see if they looked like each other when they were folded. This was exciting and I wondered what about folding up these strings from many humans . So I went to Harvard where Wally Gilbert was interested in sequencing.
At the time, very few molecular biologists were working full-time on new technology. It was considered not a suitable occupation. They said to me, you should be studying science instead of developing technology. It didn’t earn you respect. They would say things like, "well this is taking a long time, why don’t you just do an experiment, and learn about the real biology." And they would ask, "are you just playing games or do you care about the answer?"
In 1984, I published a paper called "Genomic Sequencing". That year, the US Dept of Energy had a very small meeting - about a dozen of us were there and we said we could sequence a human genome. One of the administrators at DOE saw the report and encouraged us to apply for a grant. So I did.
You must have gotten your hands dirty early on in the sequencing game.
Up until 1984, I was a grad student at Harvard. I was a post doc from 1984 to 1986 at the University of California at San Francisco. Then I went back to Harvard as a professor in 1986. As a starting professor, did a lot of experiments myself - and began to get help from students.
I was pretty sure that sequencing many human genomes would work.
I was considered optimistic by my colleagues, but now I realize how I underestimated how fast sequencing would change. The last five years it has been meteoric. Even fast technologies like computers advance by a factor of 1.5 per year, according to Moore’s Law. But genomics has improved by 10-fold per year for the last 5 years. I thought it would follow the same curve as computing. But it hasn’t.
The field of genetics depends on computers. But even though computers are moving more slowly, they are adequate.
Computers handle the sequencing information, processing trillions of bits of information, and those images are compressed and analyzed. Interpreting the genome involves more than computers. It involves new studies on many people to see how variations in genome plus environments produce our distinctive traits .
Are you excited about the 10-year anniversary?
It’s kind of an arbitrary number. For one thing, no one even sequenced an entire human genome. They made an announcement, saying it was over. They sequenced 93 percent of one genome. [People have two genomes]. They completely disregarded the issue of traits. It wasn’t even from one person. The other 7 percent is so hard, we still haven't done it.
I’m excited that we can now actually get enough genomes and trait data from volunteers so that we can work together to add value to the 10 year-old crude genome sequence that we celebrate today.
Are other countries trying to get in on this genetic race?
Beijing was a minor player in human genome 10 years ago and is a major player today. I’m working with them through my company, Knome, to sequence personal genomes of early adopters. The Chinese government has set this as a high priority. It’s like in the 60s, the U.S. decided to put a man on the moon. The U.S. was a leader in the human genome project and is now leading in genomic technology. The U.S. could launch another project, but maybe we've been waiting for the cost of sequencing to drop.
So how much does an entire genome cost these days?
The price of a genome is a little less than $10,000. The cost is $1,000. There is a gap between the price and the cost.
The gap is shrinking and the costs are still dropping.
We are going to start hearing stories about people getting their entire genome sequenced. What might these stories miss?
One thing that they might do, is try to interpret the data the same way they did for companies like 23andMe and focus on common variants.
The real story is in the rare alleles. One in ten percent of us are very affected by these, and the sequencing tests reveal them as long as we are looking for them. Almost all common diseases have a rare allele component to it. What happened before is that we went through a fad. Scientists were looking where the light was in the common variants and they mostly ignored rare alleles.
My company, Knome, can test for all 1,800 genes that have known medically actionable rare alleles. While every gene has rare alleles, a lot don’t have an impact. Clinical geneticists order these tests of rare alleles that have large and well-known impacts all the time, but just one or two tests at a time because it is so expensive. That’s going to change.
Do you think you’ll accomplish your dream soon?
I suspect everyone who wants to have their genome sequenced could have it done in the next few years. It depends on the complexity of the social interactions. Even if the price is right, it depends on who else is getting their genome sequenced. If celebrities do it, then it will become a fad and will be accepted more quickly.