That part gets left out. The tidy version is a moonshot. Thirteen years, 1990 to 2003, all of human DNA read, everybody on the White House lawn at the end of it.
The objections were not stupid. In the late 1980s, reading DNA meant Sanger's method, letter by letter, slow. At that pace three billion letters was not a project, it was a life sentence. So people said, reasonably, that it would swallow a generation's funding, that most of the genome looked like junk not worth reading, and that a cohort of good scientists would spend their careers as human sequencing machines.
It launched anyway. The first director of the US programme was James Watson, who three decades earlier had built the double helix partly on Rosalind Franklin's diffraction data, without telling her.
And the bet was never really about the sequence. It was that if you forced the cost of reading DNA down far enough to actually finish, you would build tools that changed everything afterwards. The genome was the excuse. The infrastructure was the point.
The numbers say the bet paid. NHGRI puts the first reference sequence at somewhere between 500 million and one billion dollars for the sequencing alone. By 2006, a draft human genome ran about 14 million. By late 2015, under 1,500.
That curve is most of why my industry exists. The sequencers in the labs we support are direct descendants of an argument won in 1990, against people who thought reading a whole genome was a waste of money.
Every general-purpose technology looks like a boondoggle right up until it doesn't.
I did not understand any of this when I arrived in 2016. What made it land was reading my way into Craig Venter and George Church. Venter had raced the public project with a company of his own, then set about sequencing people at industrial scale. Church had a hand in a startling number of the ventures that followed. Neither of them was chasing a single genome. They were building businesses on a ramp the Human Genome Project had bulldozed flat, and cheap.
Which brings me to the part of the Venter story that everyone gets wrong.
The version you have heard is a duel. Brash outsider takes on the plodding government project and wins. Great story. Not what happened.
Venter's company, Celera, founded in 1998, did bring a faster and cheaper approach, whole-genome shotgun sequencing, which blew the genome into millions of fragments and reassembled them by computer. The public project had been grinding along the careful, mapped, methodical way since 1990. For two years they raced, and the race genuinely drove both sides faster.
It did not end with a winner. It ended in June 2000 with a truce staged at the White House, Venter and the public project's Francis Collins standing side by side, because the only politically survivable outcome was a tie. Both sides published in early 2001, in rival journals, six days apart. Each had quietly leaned on the other's data and ideas more than either liked to admit.
The lone-maverick version flatters everyone who loves a hero and annoys everyone who was actually in the room.
That is not a separate story from the first one. It is the same one. Infrastructure does not have a single author. It cannot. The whole reason the cost curve bent is that thousands of people spent a decade doing unglamorous, repetitive, mostly uncredited work on machines, reagents, assemblers and standards, and the two competing programmes needed each other to keep going.
I see the same shape in every lab I walk into. Genome papers carry dozens of authors, sometimes hundreds. The tech who built the library. The person who sat with the sequencer overnight. The bioinformatician who turned a terabyte of noise into the one variant that mattered. They are all named on the paper.
And then, in every citation that follows, all but one of them become "et al." The orchestra is on the paper. But the Nobel has room for three.
Cost figures from the National Human Genome Research Institute. Behind this one: Siddhartha Mukherjee's The Gene, Victor McElheny's Drawing the Map of Life, Craig Venter's A Life Decoded, and James Shreeve's The Genome War.
The same argument, in a smaller room and a shorter timeframe: The sequencer is the easy part.