Thursday, September 08, 2016

Portrait Of A Genomics Instrumentation Impresario

Veteran biotech reporter Luke Timmerman's new book, Hood: A Trailblazer of the Genomics Age, is a valuable exploration of one of the leading figures in the early development of genomics and proteomics.  This in depth look at a key scientist covers not only his achievements and glories, but also his less than stellar moments and tendencies. Timmerman has combined his own interviews and research with nuggets pulled from prior news articles and oral histories of scientists who crossed paths (and sometimes figurative swords) with Hood.  While the book has issues, I would recommend it to anyone interested in the history of biotechnology.  For the full review, including numerous spoilers, read on.
Timmerman reaches back to highlight Hood's modest beginnings.  Hood's father was an engineer, but also a well known troublemaker; in his early teens Hood would take the role of designated driver to spare his family risking his inebriated father taking the wheel.  His father's profession caused frequent moves in Montana as well.  Hood's mother both burdened her son with her troubles but encouraged his talents, as did a grandfather.  Early on, Hood showed a flair for creating useful inventions.  A high school science teacher also sees promise in the young man, and fatefully, pushes hard for hood to go to school at the teacher's alma mater: Cal Tech.

Hood does well as an undergraduate and decides to go off to medical school, in part due to the plight of his younger brother with a significant developmental delay.  But he finds medical school at Johns Hopkins too focused on rote memorization, with too little exploration of the underlying phenomena of human health and disease.  For graduate school, he decides to head back to Cal Tech, linking up there with a rising immunologist named William Dreyer.  Hood and Dreyer will have a contentious relationship for many years, starting with Dreyer muscling away a project his graduate student Hood had initiated on the puzzle of antibody diversity.  Dreyer would later leave Hood off a key paper arising from that work, discounting Hood's intellectual contributions.  But Hood would also credit Dreyer for planting the idea that sometimes the best way to solve a scientific problem is to invent new tools for the job, and Dreyer also sends Hood to give talks at high-powered scientific meetings.

After completing his Cal Tech Ph.D., Hood heads to the NIH, partly because of the scientific reputation, but also as a way avoid serving as a medic in the Vietnam War.  Here, Hood's ambition shows quite prominently, as he demands a senior investigator position, skipping over being a post doc.  After the NIH, back to Cal Tech, now as an assistant professor.

At Cal Tech his lab grows, and grows, and grows.  Hood's management style becomes apparent: very little managing.  Hood also starts discovering new ways to fund his lab, such as going to donors directly, and a Farragut-esque pattern of damning the funding rules and full spend ahead.  This pattern of laissez-faire oversight, cavalier treatment of rules and wildcatting for funds will often propel him to great heights - and also generate enemies and threaten disaster.  Eventually Hood shares a Lasker Prize for his work on immune gene rearrangements with Phil Leder and Susumu Tonegawa, only to be shocked with Tonegawa is awarded a sole Nobel (could be worse; Leder has been edged out for at least 3 Nobels).  Hood eventually reaches his Cal Tech peak: selection as the chair of the Biology Department. 

It is during his Cal Tech boom years that Hood begins espousing a vision of a quartet of instruments to revolutionize biology, addressing four imperatives: sequence DNA, synthesize DNA, sequence proteins, synthesize proteins.  He first tackles the protein sequencer, which results in a royalty stream that grows his lab.  Automation of DNA synthesis is also achieved.   Later, Hood's group will develop the first automated fluorescent Sanger sequencer, leading to more funds (but more on that device later).  More on these devices - and the telling of their tales -- a bit later.  Along the way, Hood attracts talented techno-scientists such as Tim and Mike Hunkapillar.

Eventually his Cal Tech position crumbles; a revolt of the department leads to his ouster as chair. Two different scientists in his lab are discovered to have fabricated images (Timmerman leaves out the juicy bit in the original Science coverage that one consulted the other for ethical advice).  Stories abound of internal scientific warfare within his group, or of Hood not knowing the correct names of his own students or of Snakey the boa constrictor who commutes to the lab via one of the technicians.  Hood explores bolting Cal Tech, but can't find another home that really grabs him - until the Medical School Dean at University of Washington. Phil Fialkow, goes full force to recruit Hood, simultaneously snaring Bill Gates to fund the new department.  Gates and Hood become friends, happily talking biology and computers.

Hood recruits a dream team to fill out the new department, drawing both post docs from his Cal Tech group and stars plucked from other universities.  Fialkow keeps Hood happy and out of trouble.  But then the University of Washington department goes off the rails for reasons both familiar and tragic.  Hood's management style alienates key department members, while his freelance fundraising annoys the university hierarchy.  Not only funding: at one point a UW official realizes that his effort to rent space for Hood is bidding against Hood's own effort to rent the same space! An all-but-under-the-table rice sequencing effort funded by Monsanto blows up; the company pulls the funding when the genome is nearing completion, exercising the company's contractual right to hoard the data. Worse, Fialkow dies suddenly, replaced by Paul Ramsey, who has no interest in treating Hood in any special way. 

When relations with Ramsey hit the breaking point, Hood sought to have Bill Gates fund an entirely independent institute -- and receives a shocking refusal.  Hood ends up seeding his Institute of Systems Biology with an anonymous donation, anonymous because it would be embarrassing to admit it was his own money.  Hood's ISB will lurch along to the present time, through further management shakeups and funding fiascoes (and a very real fear that Hood's accounting nonchalance might have endangered the institute's non-profit status), but again with high-powered and very successful science.

Okay, that's a synopsis of much of the book - minus stuff to be discussed below and probably a lot more.  I'd wager that lousy summary is better than the Wikipedia entry on Hood, which is obviously written by a publicist.  The official biography on the ISB website is worse, never mentioning Cal Tech!  

One of the sweeter, but ultimately tragic, personal elements which Timmerman covers is Hood's wife, Valerie Logan (she would resume using her maiden name late in life), who truly complemented him.  First meeting in high school, they would date on-and-off for many years.  During his medical school stint, she gave him an ultimatum: propose or I take a job in Lebanon.  He proposed.  She endured his aggressive backpacking trips in the mountains and his long hours, but also joined in late night lab discussions.  She bore and raised a son and a daughter; Hood admits he didn't do much there.  But she also reined in some of his mistakes, such as insisting that he pay the restaurant bill he had intended to split with his grad students.  Or later, trying to help him interact productively with administrators.  Sadly, she would be stolen from him by Alzheimer's Disease.   
In addition to being stingy paying for lunches, Hood could also be stingy with credit, often giving news conferences or talks that omitted mentioning key collaborators in his lab.  This would be a point of contention with his graduate adviser Dreyer, who believes that he was the real inventor of the automated protein sequencer and Hood and students simply refined Dreyer's idea.  Hood also gained a reputation for exaggeration of his instruments' capabilities; George Church's forward to the book notes that the original "automated DNA sequencer" had essentially nothing automated.  Or exaggerating to get funds: Hood promotes an 'omics approach to prostate cancer during the early years of Michael Milken's involvement in funding prostate cancer research.  Hood proclaims that in four years he'll make significant progress against the disease, but the effort crumbles in a familiar fog of poor management.

Hood's management failings took many forms. In addition to the issues noted above, he often was away too often, delegated tasks to unreliable lieutenants (he would eventually fire three at the ISB, but only after being pushed hard by others).  In one of the most egregious bits of laziness, Hood distributes a consultant's report on the ISB's management issues to his executive team -- unconscionable since one specific recommendation is the firing of a specific individual on that team. Another example of rotten handling of an individual: Hood once promised a post doc a higher salary, then reneged.  When the post doc complained that Hood had made a commitment, Hood's reply was "Commitments should be flexible".

Timmerman generally does a good job throughout the book, but there are two general classes of issues that arise.  One is a choppy exposition of material that is apparent in two episodes, and the other is falling short of the mark on explaining the science around Hood's efforts.

The automated protein sequencer would be Hood's first big instrument splash, and also some of his first dips into serious controversy over rights, royalties and claims of inventorship.  Timmerman gives a lot of details on this, and they are valuable, but the exposition of the protein sequencer saga is weakened by being spread across two chapters, with an interlude in between.  This isn't pure adherence to a chronological storytelling method either; the restart of the story overlaps the first telling.  For example, in the first telling the existence of Applied Biosystems (ABI) is essentially a given, whereas in the second time through the details of ABI's founding are written.  That's a juicy story in itself, with Hood shopping the idea to multiple instrument makers to no avail (Beckman, founded by a legendary Cal Tech benefactor, ends up telling Hood to not darken their doorway again). Hood helps found ABI, but then reluctantly passes up on founder's shares.  

The second instance of this isn't quite as severe, but still odd.  Maynard Olson is introduced briefly and then there is a run of no mention, then he is clearly made a major player in the saga at UW.  Olson was initially enthusiastic about Hood's department, but eventually had an effective professional divorce from Hood, moving his lab as far as physically possible from Hood's facility.

Now on the science, I understand this is tricky.  Most biographies of scientists for the general public don't attempt to explain any more of the science than is completely necessary.  You won't learn how to interpret Feynman diagrams from James Gleick's Genius. Carl Djerassi even states explicitly near the beginning that he will go light on the science in his enjoyable The Pill, Pygmy Chimps and Degas' Horse. As a specialist in this field, you'd be right to guess I'd prefer a more in depth approach, but genomics geeks is far too small a market segment to seriously target.  Still, when failure to explain the science sufficiently impedes important points, I think I am on solid ground to complain.

Timmerman's description of radioactive Sanger sequencing, which Hood wished to replace with an instrument, is extremely brief - and I hope to convince you fatally brief.  

Scientists used a small jolt of electricity to push the DNA molecules through a gel. Baed on how far the DNA traveled through the gel, scientists could guess the order of its nucleotide units, A, C, G and T.  And it was a guess.  They'd peer at X-ray film images to interpret the order of the DNA sequence.  It was a long, tedious, technical process that was prone to error.

Okay, first I'll quibble with language.  "Jolt" is fast; this is a long slow tug.  And calling it a guess, particularly with a public that frequently conflates "theory" with "conjecture", is a problem as well.  It's a set of observations, and like any set of observations subject to error.  But DNA sequencing would be useless if it wasn't actually pretty accurate, and certainly more accurate than a guess.

But the more serious problem is not emphasizing how important that order is, and that any disturbance to that order is a source of errors.  Indeed, the guess language serves (in my opinion), as a red herring here, making this sound generally capricious when it is only periodically so. So, if adjacent lanes don't run uniformly (such as "smiling"), errors can occur because the bands in adjacent lanes can't be ordered correctly.  Or a "compression", in which the order of bands simply doesn't relate to increasing size.  Finally, the reduced spacing between successive bands as the fragments are longer ultimately limits the read length.  Inclusion of an actual sequencing ladder picture here would have been valuable; it is striking that Timmerman's limited photos are entirely of people, and never of any of Hood's or ABI's machines.  

Why does all this sequencing gel minutiae matter? First, by combining all four reactions in a single lane, Hood's sequencer eliminated the smiling problem.  Second, the reduced spacing between bands at long fragment lengths was still a problem; switching the design didn't help here.  But most importantly, Hood and ABI had a significant dispute over credit for inventing the sequencer.  Did Hood's lab hand a fully functional system over to ABI which the company then polished for sale, or did ABI take a proof-of-concept scheme from Hood and create an actual working sequencer?  One of ABI's claims is that the fluorescent labels developed by Lloyd Smith in Hood's lab altered the mobility of the molecules in different ways, which puts weight to the "only a proof-of-concept" view ABI put forth.  But you'll only understand that if it is made clear how important equal mobility is.

Interestingly, ABI's falling out with Hood is credited here with getting the first commercial machine to a previously unknown Vietnam medic, Craig Venter (alas, Timmerman doesn't underscore the contrast in how the two genomics giants dealt with this divisive conflict).  Hood would also get some unwelcome publicity when the Los Angeles Times alleged that federal money may have been used in the sequencer development program; if true, the U.S. government would be owed royalties.  While Hood was eventually cleared, his penchant for sloppy accounting of his grant funds had run his name through the mud (Timmerman notes that the Times never published a retraction or follow-up) and nearly caused financial disaster.

Another example of poor science reporting undermining Timmerman's story concerns one of the final straws for Hood at U of Washington.  Hood wanted to hire a surface chemist; Dean Ramsey doesn't see why a medical school should hire an esoteric chemistry specialist.  Hood says he'll just use discretionary money; Ramsey vetoes that too. I would expect a layperson to wonder whether this was tempest in teapot.  Perhaps Hood is simply becoming a megalomaniac academic, collecting one of every shiny bauble of a scientific specialist. Perhaps he is simply expanding an empire simply to feed his ego.

But those in the know will sadly shake their heads at the foolishness of a bean-counting dean completely out of his league.  How so? Because he doesn't grasp that almost all of Hood's gadgets, DNA sequencers and synthesizers, protein sequencers and synthesizers, and at this point DNA microarrays and mass spectrometers and microfluidics, all operate at (or are at least heavily influenced by) the interface between liquids and solids -- in other words, on surfaces.  The properties of these surfaces can be make-or-break, or can alter the efficiency or accuracy of one of these systems by orders of magnitude.  
Another bit of uneven coverage is Hood's involvement with starting multiple companies.  His contentious relationship with ABI gets significant ink, but that is an outlier.  Timmerman does deliver a few juicy morsels on other relationships.  Rosetta Inpharmatica fires Hood from their advisory board after he very publicly purchases a product from a direct competitor.  Hood is on the Amgen Scientific Advisory Board early on, but it is implied just as window dressing. Because Amgen took no Cal Tech intellectual property, in this case Hood is allowed to take stock -- which he profits handsomely from.  Not only that, but Hood will in later years imply he played a large role in Amgen's success. Darwin  Molecular is mentioned in passing a few times, which is disappointing; did Hood play a large role or not? Did Hood ever get deeply involved in his companies?  This is an unasked and unanswered question.

I could grouse about many more things, ranging from placing Matt Meselson at the Connecticut Bulldog Academy (his entire post-CalTech career was at Harvard), to under-reporting on Hood's group being involved in the identification of the Miller Syndrome gene by whole genome sequencing, so rare it couldn't be found by traditional genetic mapping.  Or not emphasizing that John Yates (a post doc Hood brought to UW as faculty) or Ruedi Aebersold who Hood recruited, drove protein mass spectrometry  (which is never explained at all) to be a genome-enabled replacement for Hood's Edman process protein sequencer.  Hood's lab developing an ink-jet printing strategy for making multitudes of oligos (alas, there is an extraneous "3D" tag here in the book; it really has nothing to do with 3D printing), but not how this technology has now become a key source of oligos for making DNA on a grand scale.  There's also the omission of discussion of the protein synthesizer: did Hood simply not end up making a contribution to this device or was it not covered?  

Overall, Timmerman has created a very good book with lots of memorable bits (one more: Hood was a relatively late convert to sequencing the human genome).  Hood emerges as a deeply flawed visionary, someone who could inspire great minds yet disastrously ill-equipped to manage the sorts of large science Hood was drawn to.  A likable person with many admirable qualities, so long as he remembered your name and you were delivering results or money for his research program, or were an old friend.  Someone who is scrupulously honest in terms of not lining his own pockets, but also a scientific exaggerator par excellence and one who believes that money is to be spent freely and accounted for however needed.  Lee Hood exhibits a complexity that makes him a prime subject for a biography, and now thanks to Timmerman we can all explore Hood and his impact in a balanced and comprehensive volume.

No comments: