The Violinist's Thumb: And Other Lost Tales of Love, War, and Genius, as Written by Our Genetic Code

by Sam Kean

  Then Watson, being Watson, stepped in it. In his third year as HGP director, he found out that the NIH planned to patent some genes that one of its neuroscientists had discovered. The idea of patenting genes nauseated most scientists, who argued that patent restrictions would interfere with basic research. To compound the problem, the NIH admitted it had only located the genes it wanted to patent; it had no idea what the genes did. Even scientists who supported DNA patents (like biotech executives) blanched at this revelation. They feared that the NIH was setting a terrible precedent, one that would promote the rapid discovery of genes above everything else. They foresaw a “genome grab,” where businesses would sequence and hurriedly patent any gene they found, then charge “tolls” anytime anyone used them for any purpose.

  Watson, who claimed that no one had consulted him on all this, went apoplectic, and he had a point: patenting genes could undermine the public-good arguments for the HGP, and it would certainly renew scientists’ suspicions. But instead of laying out his concerns calmly and professionally, Caligula lit into his boss at the NIH, and behind her back he told reporters the policy was moronic and destructive. A power struggle ensued, and Watson’s supervisor proved the better bureaucratic warrior: she raised a stink behind the scenes, Watson alleges, about conflicts of interest in biotech stock he owned, and continued her attempts to muzzle him. “She created conditions by which there was no way I could stay,” Watson fumed. He soon resigned.

  But not before causing more trouble. The NIH neuroscientist who’d found the genes had discovered them with an automated process that involved computers and robots and little human contribution. Watson didn’t approve of the procedure because it could identify only 90 percent of human genes, not the full set. Moreover—always a sucker for elegance—he sneered that the process lacked style and craft. In a hearing before the U.S. Senate about the patents, Watson dismissed the operation as something that “could be run by monkeys.” This didn’t exactly charm the NIH “monkey” in question, one J. Craig Venter. In fact, partly because of Watson, Venter soon became (in)famous, an international scientific villain. Yet Venter found himself quite suited to the role. And when Watson departed, the door suddenly opened for Venter, perhaps the only scientist alive who was even more polarizing, and who could dredge up even nastier feelings.

  Craig Venter started raising hell in childhood, when he’d sneak his bicycle onto airport runways to race planes (there were no fences) and then ditch the cops that chased him. In junior high, near San Francisco, he began boycotting spelling tests, and in high school, his girlfriend’s father once held a gun to Venter’s head because of the lad’s overactive Y chromosome. Later Venter shut down his high school with two days of sit-ins and marches to protest the firing of his favorite teacher—who happened to be giving Venter an F.*

  Despite a GPA well below the Mendoza Line, Venter hypnotized himself into believing he would achieve something magnificent in life, but he lacked much purpose beyond that delusion. At twenty-one, in August 1967, Venter joined a M*A*S*H-like hospital in Vietnam as a medic. Over the next year he watched hundreds of men his own age die, sometimes with his hands on them, trying to resuscitate them. The waste of lives disgusted him, and with nothing specific to live for, Venter decided to commit suicide by swimming out into the shimmering-green South China Sea until he drowned. A mile out, sea snakes surfaced around him. A shark also began thumping him with its skull, testing him as prey. As if suddenly waking up, Venter remembered thinking, What the fuck am I doing? He turned and scrambled back to shore.

  Vietnam stirred in Venter an interest in medical research, and a few years after earning a Ph.D. in physiology in 1975, he landed at the NIH. Among other research, he wanted to identify all the genes our brain cells use, but he despaired over the tedium of finding genes by hand. Salvation came when he heard about a colleague’s method of quickly identifying the messenger RNA that cells use to make proteins. Venter realized this information could reveal the underlying gene sequences, because he could reverse-transcribe the RNA into DNA. By automating the technique, he soon cut down the price for detecting each gene from $50,000 to $20, and within a few years he’d discovered a whopping 2,700 new genes.

  These were the genes the NIH tried to patent, and the brouhaha established a pattern for Venter’s career. He’d get itchy to do something grand, get irritated over slow progress, and find shortcuts. Other scientists would then denounce the work as cheating; one person compared his process for discovering genes to Sir Edmund Hillary taking a helicopter partway up Mount Everest. Whereafter Venter would strongly encourage his detractors to get bent. But his arrogance and gruffness often ended up alienating his allies, too. For these reasons, Venter’s reputation grew increasingly ugly in the 1990s: one Nobel laureate jokingly introduced himself by looking Venter up and down and saying, “I thought you were supposed to have horns.” Venter had become a sort of Paganini of genetics.

  Devil or no, Venter got results. And frustrated by the bureaucracy at the NIH, he quit in 1992 and joined an unusual hybrid organization. It had a nonprofit arm, TIGR (the Institute for Genomic Research), dedicated to pure science. It also had—an ominous sign to scientists—a very-much-for-profit arm backed by a health-care corporation and dedicated to capitalizing on that research by patenting genes. The company made Venter rich by loading him with stock, then loaded TIGR with scientific talent by raiding thirty staff members from the NIH. And true to its rebellious demeanor, once the TIGR team settled in, it spent the next few years refining “whole-genome shotgun sequencing,” a radicalized version of Sanger’s old-fashioned sequencing methods.

  The NIH consortium planned to spend its first few years and its first billion dollars constructing meticulous maps of each chromosome. That completed, scientists would divide each chromosome into segments and send each segment to different labs. Each lab would make copies of the segment and then “shotgun” them—use intense sound waves or another method to blast them into tiny, overlapping bits roughly a thousand bases long. Scientists would next sequence every bit, study how they overlapped, and piece them together into a coherent overall sequence. As observers have noted, the process was analogous to dividing a novel into chapters, then each chapter into sentences. They’d photocopy each sentence and shotgun all the copies into random phrases—“Happy families are all,” “are all alike; every unhappy,” “every unhappy family is unhappy,” and “unhappy in its own way.” They would then reconstruct each sentence based on the overlaps. Finally, the chromosome maps, like a book’s index, would tell them where their passage was situated overall.

  Venter’s team loved the shotgun but decided to skip the slow mapping step. Instead of dividing the chromosome into chapters and sentences, they wanted to blast the whole book into overlapping smithereens right away. They’d then whirlwind everything together at once by using banks of computers. The consortium had considered this whole-genome shotgun approach but had dismissed it as slapdash, prone to leaving gaps and putting segments in the wrong place. Venter, however, proclaimed that speed should trump precision in the short term; scientists needed some, any data now, he argued, more than they needed perfect data in fifteen years. And Venter had the fortune to start working in the 1990s, when computer technology exploded and made impatience almost a virtue.

  Almost—other scientists weren’t so thrilled. A few patient geneticists had been working since the 1980s to sequence the first genome of a fully living creature, a bacterium. (Sanger sequenced only viruses, which aren’t fully alive; bacteria have vastly bigger genomes.) These scientists were creeping, tortoise-like, toward finishing their genome, when in 1994 Venter’s team began scorching through the two million bases of Haemophilus influenzae, another bacterium. Partway through the process, Venter applied for NIH funds to support the work; months later, he received a pink rejection notice, denying him money because of the “impossible” technique he proposed using. Venter laughed; his genome was 90 percent done. And soon afterward the hare won the
race: TIGR blew by its poky rivals and published its genome just one year after starting. TIGR completed another full bacterium sequence, of Mycoplasma genitalium, just months later. Ever cocky, Venter not only gloated about finishing both first—and without a red cent from the NIH—he also printed up T-shirts for the second triumph that read I MY GENITALIUM.

  However begrudgingly impressed, HGP scientists had doubts, sensible doubts, that what worked for bacterial DNA would work for the far more complicated human genome. The government consortium wanted to piece together a “composite” genome—a mishmash of multiple men’s and women’s DNA that would average out their differences and define a Platonic ideal for each chromosome. The consortium felt that only a cautious, sentence-by-sentence approach could sort through all the distracting repeats, palindromes, and inversions in human DNA and achieve that ideal. But microprocessors and sequencers kept getting speedier, and Venter gambled that if his team gathered enough data and let the computers churn, it could beat the consortium. To give due credit, Venter didn’t invent shotgunning or write the crucial computer algorithms that pieced sequences together. But he had the hubris (or chutzpah—pick your word) to ignore his distinguished detractors and plunge forward.

  And boy did he. In May 1998, Venter announced that he’d cofounded a new company to more or less destroy the international consortium. Specifically, he planned to sequence the human genome in three years—four years before the consortium would finish—and for one-tenth of its $3 billion budget. (Venter’s team threw the plans together so quickly the new company had no name; it became Celera.) To get going, Celera’s parent corporation would supply it with hundreds of $300,000, state-of-the-science sequencers, machines that (although monkeys could probably run them) gave Venter more sequencing power than the rest of the world combined. Celera would also build the world’s largest nonmilitary supercomputer to process data. As a last gibe, even though his work threatened to make them superfluous, Venter suggested to consortium leaders that they could still find valuable work to do. Like sequencing mice.

  Venter’s challenge demoralized the public consortium. Watson compared Venter to Hitler invading Poland, and most HGP scientists feared they’d fare about as well. Despite their head start, it didn’t seem implausible that Venter could catch and pass them. To appease its scientists’ demands for independence, the consortium had farmed its sequencing out to multiple U.S. universities and had formed partnerships with labs in Germany, Japan, and Great Britain. With the project so scattered, even some insiders believed the HGP satellites would never finish on time: by 1998, the eighth of the HGP’s fifteen years, the groups had collectively sequenced just 4 percent of human DNA. U.S. scientists were especially trembling. Five years earlier, Congress had eighty-sixed the Superconducting Super Collider, a massive particle accelerator in Texas, after delays and overruns had bloated its budget by billions of dollars. The HGP seemed similarly vulnerable.

  Key HGP scientists, however, refused to cower. Francis Collins took over the consortium after Watson’s resignation, albeit over the objection of some scientists. Collins had done fundamental genetics work at the University of Michigan; he’d found the DNA responsible for cystic fibrosis and Huntington’s disease and had consulted on the Lincoln DNA project. He was also fervently Christian, and some regarded him as “ideologically unsound.” (After receiving the consortium job offer, Collins spent an afternoon praying in a chapel, seeking Jesus’s guidance. Jesus said go for it.) It didn’t help matters that, in contrast to the flamboyant Venter, Collins seemed dowdy, once described as having “home-cut hair [and a] Ned Flanders mustache.” Collins nevertheless proved politically adept. Right after Venter announced his plans, Collins found himself on a flight with one of Venter’s bosses at Celera’s money-hungry parent corporation. Thirty thousand feet up, Collins bent the boss’s ear, and by the time they landed, Collins had sweet-talked him into supplying the same fancy sequencers to government labs. This pissed Venter off no end. Then, to reassure Congress, Collins announced that the consortium would make the changes necessary to finish the full sequence two years early. It would also release a “rough draft” by 2001. This all sounded grand, but in practical terms, the new timetable forced Collins to eliminate many slower satellite programs, cutting them out of the historic project entirely. (One axed scientist complained of “being treated with K-Y jelly by the NIH” before being you-know-whated guess-where.)

  Collins’s burly, bearded British counterpart in the consortium was John Sulston, a Cambridge man who’d helped sequence the first animal genome, a worm’s. (Sulston was also the sperm donor whose DNA appeared in the supposedly realistic portrait in London.) For most of his career, Sulston had been a lab rat—apolitical, and happiest when holed up indoors and fussing with equipment. But in the mid-1990s, the company that supplied his DNA sequencers began meddling with his experiments, denying Sulston access to raw data files unless he purchased an expensive key, and arguing that it, the company, had the right to analyze Sulston’s data, possibly for commercial purposes. In response Sulston hacked the sequencers’ software and rewrote their code, cutting the company off. From that moment on, he’d grown wary of business interests and became an absolutist on the need for scientists to exchange DNA data freely. His views became influential when Sulston found himself running one of the consortium’s multimillion-dollar labs at the (Fred) Sanger Centre in England. Celera’s parent corporation happened to be the same company he’d tangled with before about data, and Sulston viewed Celera itself as Mammon incarnate, certain to hold DNA data hostage and charge researchers exorbitant fees to peruse it. Upon hearing Venter’s announcement, Sulston roused his fellow scientists with a veritable St. Crispin’s Day speech at a conference. He climaxed by announcing that his institute would double its funding to fight Venter. His troops huzzahed and stomped their feet.

  And so it began: Venter versus the consortium. A furious scientific competition, but a peculiar one. Winning was less about insight, reasoning, craft—the traditional criteria of good science—and more about who had the brute horsepower to work faster. Mental stamina was also critical, since the genome competition had, one scientist noted, “all the psychological ingredients of a war.” There was an arms race. Each team spent tens of millions to scale up its sequencing power. There was subterfuge. At one point two consortium scientists reviewed for a magazine the fancy new sequencers Celera was using. They gave them a decidedly mixed review—but meanwhile their bosses were secretly negotiating to buy dozens of the machines for themselves. There was intimidation. Some third-party scientists received warnings about their careers being over if they collaborated with Venter, and Venter claims the consortium tried to block publication of his work. There was tension among purported allies. Venter got into innumerable fights with his managers, and a German scientist at one consortium meeting screamed hysterically at Japanese colleagues for making mistakes. There was propaganda. Venter and Celera crowed their every achievement, but whenever they did, Collins would dismiss their “Mad magazine” genome, or Sulston would appear on television to argue that Celera had pulled another “con.” There was even talk of munitions. After employees received death threats from Luddites, Celera cut down trees near its corporate campus to prevent snipers from nesting in them, and the FBI warned Venter to scan his mail in case a Unabomber wannabe targeted him.

  Naturally, the nastiness of the competition titillated the public and monopolized its attention. But all the while, work of real scientific value was emerging. Under continued criticism, Celera felt it once again had to prove that whole-genome shotgunning worked. So it laid aside its human genome aspirations and in 1999 began sequencing (in collaboration with an NIH-funded team at the University of California, Berkeley) the 120 million bases of the fruit fly genome. To the surprise of many, they produced an absolute beaut: at a meeting just after Celera finished, Drosophila scientists gave Venter a standing ovation. And once both teams ramped up their human genome work, the pace was breathtaking. There
were still disputes, naturally. When Celera claimed it had surpassed one billion bases, the consortium rejected the claim because Celera (to protect its business interests) didn’t release the data for scientists to check. One month later, the consortium itself bragged that it surpassed a billion bases; four months after, it preened over passing two billion. But the harping couldn’t diminish the real point: that in just months, scientists had sequenced more DNA, way more, than in the previous two decades combined. Geneticists had excoriated Venter during his NIH days for churning out genetic information without knowing the function. But everyone was playing Venter’s game now: blitzkrieg sequencing.

  Other valuable insights came when scientists started analyzing all that sequence data, even preliminarily. For one, humans had an awful lot of DNA that looked microbial, a stunning possibility. What’s more, we didn’t seem to have enough genes. Before the HGP most scientists estimated that, based on the complexity of humans, we had 100,000 genes. In private, Venter remembers a few straying as high as 300,000. But as the consortium and Celera rifled through the genome, that estimate dropped to 90,000, then 70,000, then 50,000—and kept sinking. During the early days of sequencing, 165 scientists had set up a pool with a $1,200 pot for whoever came closest to guessing the correct number of human genes. Usually the entries in a bubble-gum-counting contest like this cluster in a bell curve around the correct answer. Not so with the gene sweepstakes: with every passing day the low guesses looked like the smartest bets.