Tuesday, November 15, 2016

HGP Counterfactuals, Part 5: HGP Stifled

Okay, enough procrastination. I've outlined the general idea of counterfactuals and then obsessively detailed the generation of physical maps, planning the sequencing of BAC clones and the rapid winnowing of sequencing technologies during the early stages of the genome project.  Time for a main act: what if the public Human Genome Project never happened?

The idea of a Human Genome Project was met with more than a small amount of derision and opposition.  Sydney Brenner, no slouch in clever genomics ideas, suggested that those convicted of scientific misconduct be sentenced to the drudgery of running sequencing gels. Many feared the growth of gigantic genome centers (which happened), worrying that they would not go away when the genome was complete (they didn't). Others saw little value.  And probably a few dozen other reasons; I'm no scholar on the topic.  These individuals argued against the project and fought it; what if they had succeeded?

Now, I realize one bit of difficulty with this, bordering on arrogance to pursue it.  First, by no HGP I don't mean an HGP put off for a few years, but completely blocked.  That would require the opponents to successfully fight off what would have almost certainly been a recurring attempt by HGP proponents to launch the project.  Many fun counterfactuals work off some binary, definitive event -- what if Churchill had died of natural causes in the 1930s or the Lost Order hadn't been lost.

Second problem is assuming the HGP could have been shut down everywhere.  The U.S. contributed a lot, but so did many other nations -- I'm sure to miss some key ones if I attempt to enumerate all.  So this scenario can be seen as a rejection in the U.S. causing other countries to similarly hold off, which perhaps isn't realistic.  Perhaps U.S. abdication of a role would have triggered China or Japan to invest more, driving the project further but with a different geographical distribution of the investment and work.

But given these caveats, let's try to imagine.  Some sort of definitive stop comes to the HGP effort; perhaps a key Congressman (such as the Speaker of the U.S. House or the Senate Majority Leader) declares that they are sick of the proposal and won't let it on the floor again.  Or the President vetoes, taking the wind out of the sails.  HGP is dead.

That wouldn't mean human genome research or sequencing technology efforts would shut down in the U.S., just that there wouldn't be a huge directed funding effort.  No HGRI.  Without a field awash in funds, no lavish Hilton Head meetings with small mountains of lobster and snow crab served by the pool (ABI paid for that, but of course that was off the sequencing revenues).  

This is where those network effects I mentioned yesterday start being important.  ABI and the others still sell plenty of ABI sequencers, but not on the same scale.  Most are in small core facilities which have an instrument or two, not huge fleets in sequencing centers. 

Even with this diminution, EST sequencing would be expected to take off.  The idea of sequencing random cDNA clones originated with Gregor Sutcliffe and colleagues, but it was Craig Venter who expanded the idea to a grand scale and grand goal.  So in this timeline, perhaps on about the same schedule, Human Genome Sciences forms and later begats Venter's TIGR (in a way). Other companies such as Incyte and Curagen jump on the bandwagon as well.  Merck still pays to generate a fully public counterweight to HGS.

If you doubt this, remember that Walter Gilbert came very close to launching a genome sequencing company in the late 1980s; he told me that this was thwarted only by the 1987 stock market crash.  Genome Corp had an initial SAB that was a nearly complete list of the genomics stars of the times; I doubt it would have held together, but it was an impressive list (though if memory serves, Venter wasn't on it). So there was appetite in the investment community for such companies.  But would a public rejection of an HGP been seen as encouraging (not competing against free!) or negative (if they don't want it, why should we?).

The rejection of a public HGP might well have slowed investment in the early positional cloning companies.  That was Millennium's original guise, that they would use genetic mapping to identify key drug targets.  Sequana was another competitor and Mercator yet another with a similar business plan.  The public genome project, by funding continuing improvement of genetic and physical map in reality helped them, though the physical maps mostly showed up after these business models had gone out of favor.

What about the model organism projects?  These were to some degree marketed as pilot projects for the human genome effort, though as I noted yesterday that these practical projects may have inhibited technology development.  As smaller genomes of important models, there tended to be less opposition to investing in their sequences. But in this scenario, I'll argue they suffer the same fate as human: sequencing is limited to what individual labs can get funded via R01 and similar mechanisms.  Alternatively, C.elegans sequencing fires up at the Sanger Centre, once that is funded via the Wellcome Trust's proceeds from the Glaxo Wellcome merger. 

Still, in this scenario Venter's track through history is little changed.  His sequencing operation grows in efficiency more slowly, due to much smaller network effects on the ABI platform.  George Church's lab continues to refine the multiplex sequencing method, using it to sequence cosmids and BACs.  Because the method can work with a very small team, it continues forward.

So it is perhaps a few years later than the real timeline, around 1997 or 1998, when Craig Venter stuns the world by generating the complete genome sequence of a bacterium.  As with ESTs, he wasn't the first to think of it; Church had started sequencing two bacteria in parallel with the multiplex method before deciding further improvements were needed first.  But in this alternative world, as in the real one, Venter's gifts shine through, that is the gifts for attracting bright minds around him and for recognizing when a technology can be pushed hard to a goal.  Venter's success attracts investors to go for the whole enchilada. So Celera is formed a few years off the real timeline and delivers to its corporate customers a human genome around 2002.  The genome is very fragmentary and most pieces are adrift rather than anchored to a specific genomic location.  But the customers don't care; they wish only to have privileged access to the scope of possible drug targets.

This leads to another key question.  Suppose Celera were to launch in a world without an HGP; would this suddenly shock the U.S. and others into launching a last-ditch effort?  Or would it be seen as a lost race, with perhaps access to Celera's database negotiated for the NIH, Howard Hughes Medical Institute or other large organizations -- but no human genome at the public's fingertips.  

Now we are getting farther away from initial events, with the contingencies and interactions multiply.  No public HGP, no HGRI and no genome centers probably means no HapMap and no big era of GWAS.  Without streams of SNPs, array companies such as Affymetrix and Illumina stick to RNA profiling.  

What would have happened to non-electrophoretic methods in this scenario?  To what would have become known as "next-gen" sequencing?  Many of these germinated from ideas broached in the early 1990s -- along with a lot of duds.  I'll scan those a bit more next installment, but it is noteworthy that the earliest efforts came from private companies (Curagen/454, Lynx, Manteia, Solexa) or genome-focused small labs (polonies from Church, Helicos from Quake), not genome centers.  Genome centers played a role (particularly the Sanger Centre with Solexa), at least in my weak understanding (I should consult Kevin Davies' $1000 Genome book).  Genome centers did a lot of the beta testing on these systems, but the rise of the Polonator purely from small groups suggests that single molecule and polony sequencers might still have emerged.  However, some of the support for these was in a follow-on to the HGP which supported advanced sequencing technologies; with no HGP this support might not have emerged.  So perhaps advanced sequencing would not have hit until as late as the beginning of this decade.

One other important ripple effect of a no-HGP scenario: hundreds and hundreds of scientists not trained in genomics.  Many, many clever engineers, mathematicians, computer scientists and the like not lured into biology.  

I've strayed farther into this alternate future than perhaps is wise, but I think it has been an interesting exploration.  I truly didn't know where it would take me when I started pondering it. Only after some toying with ideas did it occur to me that Venter's track was largely, though not fully, independent of the public genome project.  The genome would have been available to humanity, but only by paying a fee to Celera, rather than free on Internet for all.  The genome would likely have been significantly less useful as well, a set of sequence islands unconnected to the maps which human geneticists traffic in. So in terms of ROI, science gets a human genome for very little, but also gets far less value from it.  

Tomorrow, my second alternative scenario: HGP proponents "settle" for a methods-focused project on a tight budget.

No comments: