Is Silicon Valley ready for biotech’s newest unethical prodigies?
The future was looking unbelievably bright for Forbes 30-under-30 healthcare golden boy Gabriel Otte, CEO of liquid biopsy startup Freenome, the recipient of $65M in seed funding from Silicon Valley venture capital firm Andreessen Horowitz, also known as a16z. A Cornell graduate who worked at Apple at 17, Otte enrolled as a graduate student at the University of Pennsylvania in 2011, working in the lab of Dr. Shelley Berger within the Genomics and Computational Biology (GCB) program. Otte then left prematurely under somewhat murky circumstances — according to him, “[on leave because] starting a company violated my PhD contract” (link).
Regardless of the reasons for his premature departure, this setback did not stop Otte from outright claiming to have completed his PhD in interviews, listing himself “Gabriel Otte, PhD” on his company’s website (though this was recently scoured) and full billing as a headliner at industry conferences. Depending on who you talk to in scientific circles, faking these credentials ranges on the “not right” spectrum from just a clear-cut no-no to an egregious ethical violation.
My role in this tale is quite small and tangential. I’m not a PhD and I’ve never met Gabriel Otte — I’m just a bioinformatics programmer and overripe doctoral student who was trying to find someone’s scholarly output. Granted, I was especially curious as to how a genomics entrepreneur and CEO could have no first-author or second-author papers, and no open source bioinformatics projects. Except for some minor placings on Berger lab manuscripts, he was more or less unknown in science. Otte claimed to have “published several papers on applications of machine learning on large genomics datasets and other big data”. These don’t exist. He was missing from the GCB alumni page. Penn makes all dissertations in the past 10 years publicly available. His wasn’t.
Having compiled this scant evidence I contacted Buzzfeed senior technology reporter Stephanie Lee, who was actively writing about Freenome. Lee previously worked at the San Francisco Chronicle, and now wears two hats as “industry acolyte” and “investigative reporter”, but her portfolio shows she is not uncomfortable in the latter role. Lee pieced together some details on Otte’s mysterious exit from the graduate program, and a host of other unfortunate typos and misunderstandings that have plagued him. And so, like many compelling stories that involve a bigger truth, this one began with the journalist’s lead fighting to cover up a small lie that really wasn’t even worth telling.
If you haven’t read Lee’s article yet, stop right now and read it here. It is some wild stuff — really just bizarre.
I’m not sure if this is an isolated incident, part of a bigger pattern of deception, or just a minor stumbling block in Otte’s meteoric rise, but it has made me think about youth, ethics, and having no firsthand Silicon Valley startup experience myself, what I can only surmise is some kind of risk-taking hero-worship that dispenses with the most basic of professional vetting standards in exchange for a good song-and-dance. This story of Otte detecting cancer in the prostate of his own father, a distinguished philosophy professor at UC Santa Cruz, is looking rather more dubious by the day. Does your father call you on the phone to report that his PSA results are normal, a test notorious for its false positives? Mine calls to discuss home improvement projects. Doubting this feel-good father-son story doesn’t give me pleasure, because there is good reason to believe liquid biopsies of cell-free DNA will be more predictive and less invasive than standard protocols.
Surely this young man could have claimed to have dropped out for any number of reasons — no one would have questioned it, or even cared. Buoyed by the enduring mythology of technology companies founded by grad, and even undergrad, dropouts, there is no shame in leaving early. This story has led to endless speculation in my group of friends and colleagues, many of whom keep asking me the same question: Why would Otte take such a risk in misrepresenting himself about something that is so easy to disprove?
We can’t know if Otte is misunderstood, a pathological liar, or just young and foolish. The grown-up at Freenome is supposed to be Andreessen Horowitz general partner Vijay Pande, PhD. Dr. Pande is a giant in the field of computational biology, and himself is accustomed to the playing the role of celebrated young phenom. But either he didn’t care, or he was so entirely fixated on an opportunity to use AI in cancer so as to not even bother to call for a reference. With millions at stake, where was the due diligence?
Even with a brilliant guiding hand, is it a good idea to have someone in their 20’s running a biotech? Case in point: Elizabeth Holmes was 19 when she started Theranos. Science is not engineering, and building clinical diagnostics is very different from building mobile apps — there are too many avenues to engage in “p-hacking” and other forms of selective truth, even subconsciously. The best scientists I’ve worked with tend be confident in their abilities but initially skeptical of their own results. It takes a long time to learn to do that instinctively, even without having to manage the expectations of investors, the media, and peanut gallery hecklers like myself. I would assume Silicon Valley venture capital firms are looking for a clean “exit”, not pedantic lessons in false discovery rates and reproducible research. Placing a twenty-something, PhD or not, as the primary advocate for caution seems like a recipe for disaster.
All this suggests the “fake it ’til you make it” startup culture is getting worse. This is a dangerous time for that, because in healthcare and biotech there is an increasing emphasis on predictive exotic black-box machine learning techniques of the sort Freenome is developing. That’s ok — either you predict cancer early or you don’t — but the mechanisms underlying these predictions will be all but inscrutable by old-school biostatisticians at the FDA. It’s not clear how these type of diagnostics, should they learn and improve over time, will fit into how pharmaceuticals have been tested for the past century. It’s also not clear how these type of diagnostic startups are not ripe for top-down fraud — the kind that rears its head long after VCs make their exit. In the wake of Theranos, I would have thought someone in charge would start tightening the screws. I guess I’ll just keep searching for dissertations.
