Here Is a Human Being (31 page)

P
aper Japanese shades were suspended by long strings over the sliding glass doors in the back of George and Ting’s house, which meant they covered only the bottom part; early winter morning sun streamed through the top. No overhead lights were on; the light in the room was natural except for the heat lamps under the table against the wall where the family turtles basked (as a child, Marie was an avid turtle breeder and it remains an abiding Church-Wu pastime). Out in the backyard the pond was partially frozen, the mulberry trees were bare, and snow was everywhere. It was barely 7:15 but Ting was already gone to the lab to tend to her mutant fruit flies.

George, who was waiting at the door when I arrived, had been up for hours. He described himself as “in sync with the sun” and “very circadian,” although in fact he was more than that: he was narcoleptic.
¹
At one point during our first big PGP event, as most of the PGP-10 sat around the table with cameras rolling, he fell asleep. For a moment it was almost imperceptible: he just sat back in his chair, head still upright, eyes closed. The rest of us stole furtive glances at one another … should we wake him? Ignore him? In a couple of minutes he was back as though nothing had happened. “The big disadvantage of narcolepsy is the social stigma,” he told me. “People get insulted when I fall asleep in the middle of a one-on-one conversation. And it does compromise me somewhat: I always miss some part of lab meeting. People can wake me if they want, but they usually don’t. That’s just how our society is set up.

“My daughter is just like me. What’s really funny is when she and I go to the sleep clinic with my wife. Marie and I will both be asleep halfway through the conversation,” he said.
²
“It’s the same at night: I tell Ting that I’m so keyed up I don’t think I’ll be able to sleep—and then boom, before my head hits the pillow I’m out.”
³

“We do want to figure out what’s going on with our sleep, especially with Marie, since she’s trying to get through school. I already did the flunk-out thing; we want to try to keep that to a minimum in the family.”
⁴

He filled a coffee cup with water at the sink and handed it to me. His relationship to food was as idiosyncratic and indifferent as his relationship to sleep. I asked him whether he chose not to eat animal products for moral or dietary reasons. “Neither,” he said. “For biochemical ones.” He ate once a day. He claimed to like the taste of food but, as far as I could tell, never experienced deep pangs of hunger. Yet at 245 pounds he was hardly wasting away. “I can go for days without eating,” he shrugged. “Or I can eat to excess.”
⁵
I got used to seeing him at meals where he was a bored spectator while those around him cleaned their plates with gusto.

We sat in the corner of the Church family room/dining room, him at his computer desk and me on the couch behind him. He was working on a manuscript and, in doing so, indulging his other scientific passion: synthetic biology. One of his long-term objectives was to engineer a genome from scratch. Craig Venter’s team had already taken a bacterium and systematically removed its genes one by one in order to identify the minimum number needed to sustain life. Venter hoped to synthesize the first man-made bacterium,
Mycoplasma laboratorium.
⁶
But George was undertaking something different: he wanted to tinker with the genetic code itself—that is, to rewrite it so that the three-letter DNA “words” would code for different amino acids than they do in nature. One could see the geek appeal of this, especially to an inventor: make something completely new and build new parts with which to do it. That said, one could imagine practical applications for invented genomes. If crops were to use a novel genetic code, for example, then they would be of no use to pathogens: every such crop would immediately become virus-resistant, even to viruses that had yet to be discovered. Jim Watson had said more than once that “someone has to play God,”
⁷
and indeed, he often volunteered; but George was always a bit more reticent. “It’s kind of weird,” he said, “this idea of making something impervious to viruses we know nothing about.”
⁸

As ever, he was driven by advances in sequencing technology, and as ever, his drive seemed to me untainted by any of the Buddhists’ Three Poisons: ignorance, hatred, or desire.
⁹
When I asked how important it was that his lab be the one to win the X Prize, his response was basically “not at all.” He was nominally affiliated with almost all of the seven registered competitors in the contest to sequence a hundred genomes in ten days for less than $1 million, though only a victory by his “Personal Genome X-team” and its fleet of Polonators could actually bring him the glory. He insisted it didn’t matter. “I am totally unconflicted. I hope everybody wins.” The bottom line, he said, would be the bottom line: whoever could sequence genomes on the cheap would get the brass ring. Illumina had become the dominant next-generation sequencing platform, in George’s lab and everywhere else. “The most cost-effective one and the one with the longest reads [among the short-read platforms] is Illumina,” George said. “We use Illumina for almost everything. It is the one to beat or improve upon.”
¹⁰

But he hoped that that improvement would happen soon. When, in 2008, Complete Genomics announced that it would begin sequencing whole human genomes for five thousand dollars in the coming year, George practically clapped his hands with joy. “I’m delighted! This changes the game. Since 2004 sequencing has been getting ten-fold cheaper every year. Can we squeak out one more factor-of-ten reduction in one year? I sort of doubt it … but that would be pretty exciting! My principle is this: I don’t want to sequence my genome or anyone else’s until the price is right. Now I think the price is right. And nobody’s suing anybody! Isn’t that great?”
¹¹

For a moment, it was. But Complete Genomics got blindsided by the financial collapse just like everyone else—the retail five-thousand-dollar genome would have to wait until at least 2010. Meanwhile, in 2009 Life Technologies (the entity formed by the merger of ABI and Invitrogen
¹²
) sued Illumina, which promptly countersued.
¹³

As a young scientist in 1987, a few years after Tito’s death during the waning days of the Socialist Federal Republic of Yugoslavia, Rade Drmanac (RAH-day Dur-MAHN-itch) submitted what he assumed would be a doomed grant application to the U.S. Department of Energy proposing a novel approach to DNA sequencing. He was a young guy from the wrong side of the Iron Curtain petitioning Reagan’s America for research funds. But to his utter shock, DOE loved it—he got the money. “One hundred fifty thousand in real U.S. dollars!” he remembered. “In a communist country and from the same DOE that builds atomic bombs!” Among the swooning reviewers of his proposal was another young scientist, who wrote, “This is the best theory ever proposed for sequencing a human genome.” Drmanac told me that even though his then-brilliant idea was premature (the necessary camera and computer technology did not yet exist), he still recalls George Church’s lofty praise with pride.
¹⁴

It turned out that getting the grant was the easy part. Drmanac worried that Yugoslav officials would keep the DOE’s dollars and pay the award in dinars, whose value was diminishing by the day; the country was on the brink of hyperinflation. There was also concern that the government would skim some off the top. The solution, Drmanac and his fellow scientists reasoned, was to keep the money on U.S. soil. “We opened an account in the U.S. embassy in Belgrade,” he said. “So the dollars never went out of there. We would bring in bills and invoices and embassy officials would pay them from our grant. Not a single dollar was lost.”
¹⁵

Drmanac has a stout profile, a thicket of gray and black curly hair, a boxer’s nose, and a ready smile that accents his crow’s-feet. When he’s not wearing them, his glasses are hooked into the V-neck of his sweater. As Yugoslavia continued its inexorable disintegration, Drmanac moved to the United States and worked at Argonne National Laboratory in suburban Chicago for a while before starting his first company. He doggedly pursued DNA sequencing technology until the Human Genome Project caught fire and then switched to gene
discovery,
which was all the rage, but something he considered to be a distraction—"the gold rush,” he called it. Once the first complete (and expensive) human sequence was in hand and it became clear that finding genome-based drugs was not going to be a cakewalk, the investment climate for DNA sequencing technology became hospitable again: we needed more genomes if we were ever going to understand them. Around this time Drmanac met Cliff Reid, an easygoing software entrepreneur who happened to have degrees from MIT, Harvard, and Stanford. The two crafted Drmanac’s technology into a business plan and set out with a simple but tantalizing message for VCs: “We are trying to achieve something other sequencing companies are planning to do in ten years. We are not interested in releasing a product that is only ten percent better than what’s already out there. If we cannot develop a method in two years’ time to sequence a complete human genome for less than $10,000, then we will have failed.” By 2007 Drmanac and Reid’s new company, Complete Genomics, had raised more than $12 million—not nearly enough, but enough of a vote of confidence to keep going.
¹⁶

Not only did the company have new technology and new money, it had a novel business model. Unlike 454, Solexa, ABI, Helicos, and a smattering of other would-be players in the suddenly wide-open sequencing space, Complete Genomics did not want to get into the instrument business. “Market penetration for new machines is
so
slow,” said Drmanac. “Early on we realized that even if we had an instrument
today,
people would not be able to use it.” Or want to, necessarily. In a previous company he’d started, Drmanac had developed a small chip upon which one could sequence several thousand base pairs of DNA—a few genes’ worth, which was a lot at the time. His thinking then was that every lab would have one on every bench. “But when we started talking to people, they said, ‘No, we prefer to send our samples out to be sequenced.’ ”
¹⁷

Many if not most life scientists believed that their time was better spent actually
doing science
than learning how to operate a new machine that was likely to become obsolete in a couple of years. Pharmaceutical companies, for example, mired in a long slump and hit hard by the economic downturn, had made it clear that certain expenditures were on the chopping block: half-million-dollar toys would no longer be purchased with impunity.
¹⁸
Nor were would-be customers excited about the massive quantities of ancillary data that were generated with every complete sequence.
¹⁹
Complete Genomics’ DNA sequencing product would therefore be a black box: DNA samples in, sequence out. And the company would store genomic data for customers at its headquarters. It was a pure fee-for-service play: genomes at wholesale.
²⁰

By 2008 investors and board members were anxious. Complete Genomics had been operating in stealth mode for months, which can feel like a long time in Silicon Valley. Company insiders pressured Reid and Drmanac to demonstrate proof of principle to the world: “Sequence
E. coli,”
they urged. The two founders resisted. “Success in bacteria doesn’t guarantee success in the human,” said Drmanac. “From day one we were going to be sequencing the
human genome!
I was tired of sequencing small organisms. I said, ‘Human genome. Let’s do it.’ ”
²¹

In July 2008 they did it, and reportedly for less than four thousand dollars in materials.
²²
Eschewing peer review, a few months later they revealed many of the details on their own website. Drmanac described it to me in simple terms. “Take human genomic DNA, randomly fragment it, make arrays, hybridize, ligate, and sequence.” The arrays were regular microscope slides with a billion tiny slots, each of which could accommodate a single wad of balled-up DNA: a nanoball. Unlike other next-gen technologies, it required no beads or other substrates: just pure DNA. A single slide could yield seventy gigabases of DNA, or the sequence of a single human genome twenty-three times over, and with an error rate of less than 0.1 percent.
²³
By September 2009, Complete Genomics had sequenced fourteen whole human genomes.
²⁴

When I visited the company’s headquarters it looked to be typical of Redwood Shores, California: an angular mix of salmon-colored brick and mirrored glass surrounded by lush greenery. It sat just beyond the fringe of the Google campus. Inside was a sort of chaotic feng shui: a trickling fountain drowned out by leaf blowers and lawn mowers outside the windows, cushy furniture, boxes piled up in the hallways. The kitchen was lavishly appointed with free drinks, multiple microwaves, and two refrigerators. Instead of an office, Drmanac had a cube—a big cube, but still a cube. There was a “dot-commie” feel to the place. “I think it’s a very healthy culture there,” said an admittedly biased George. “It’s one of innovation and an amazing amount of openness, especially considering their business model is as closed as it could possibly be.”
²⁵
The marketing director was wary of me, which was probably appropriate. She gave me a cursory tour of the labs, seemed a bit impatient with my questions, and insisted I meet with the company’s legal counsel to sign a nondisclosure agreement. We were a long way from the Polonator.

Despite its positive buzz and relatively low overhead, Complete Genomics was hit by the same freight train as nearly every other start-up enterprise in the early part of 2009: extreme investor skittishness and virtually no money to be had. The company cut salaries. Its big ambitions—one thousand genomes in 2009 and twenty thousand in 2010—had to be scaled down in the wake of the global financial meltdown. “Our timing could not have been worse,” CEO Cliff Reid told
In Sequence.
“We started this [round of] financing the day that Lehman Brothers failed.”
²⁶