Shopping for Your Genome
I spent the last couple days at the Personalized Medicine World Conference and one of the most intriguing companies I encountered was a small startup called HolGenTech, which aims to combine genetic data and mobile devices to give people just-in-time recommendations about what they should and shouldn't consume. It's an interesting concept--nutrigenomics meets mobile computing--but I'm skeptical that, in the near-term, genetic research will be able to offer the sort of value the company claims it will offer to the end user. At the same time, their effort to rush to market highlights the gulf between what consumers and entrepreneurs hope breakthroughs in genetics (and neuroscience) will offer and what may actually be scientifically valid.
In brief, HolGenTech's service works as follows:
The PGA [Personal Genome Assistant--marketing speak for a smart phone] uses a device’s bar code reader to capture product ingredient information and respond with personalized screens of recommendation advice and ratings that display on a scale of -10 to +10, corresponding to analysis of integrated data from multiple sources. The PGA user can automatically and immediately identify the personalized prevention efficacy of any product under consideration, as long as the product has a bar code for ingredients. Consumers are equipped to make quick, yet thorough, product comparisons that take into consideration personal health preferences and genomic information with special attention to a disease, syndrome, or health condition they wish to improve.
And this sounds great--and extremely useful. In their demo video, however, they inadvertently showcase just how far they are from delivering on this claim. Using a DNA sample from a writer, they note that she is lactose intolerant and then go onto show how, by taking pictures of different product barcodes, she can be told which products might and might not be problematic for this condition. It's not at all clear why taking a picture of a barcode would be simpler than reading a list of ingredients--something people with food allergies do quickly all the time--or, for that matter, why someone would need a genetic scan to know that they have lactose intolerance issues.
The reason, I imagine, that these recommendations fall so flat is that conclusions in genetic research remains frustratingly vague and inconclusive, despite huge advances in genetic sequencing and analysis. In a recent paper in Nature (subscription required), a group of genetics researchers found that direct-to-consumer genetic testing companies gave mostly contradictory results. The testing services often disagreed about what one's genes might mean, in a general sense, about whether her lifetime risk of developing a disease might be above or below average based on her genetics. In that context, delivering daily, well-supported recommendations about specific products and services seems like a pipe-dream.
So why am I intrigued by a pipe-dream? Because HolGenTech represents what we hope genetics can give us. Genetics, neuroscience, and several other fields in health and health care are making hugely fast advances, in terms of basic science research. And that research is yielding complex, intriguing findings--most of which will take a decade or more to move into mainstream products and services.
It's this latter point--speeding the pace of research getting out of laboratory settings and into the hands of everyday people--that I think will form the core of an enormous challenge in health and health care in the coming decade. Because I do think there's demand to try to apply all of this research in services like HolGenTech--even if no one yet knows how to do it in a thoughtful way.
- Bradley Kreit's blog
- Login to post comments
-
A Response from HolGenTech's Founder
Bradley Kreit
HolGenTech's founder, Andras Pellionisz, emailed me a fairly extensive response to my blog post, which I've reproduced below.
A couple quick points: Andras implies, a couple times here, that I didn't watch HolGenTech's demo; I reviewed it carefully almost immediately before writing the post above (hence my direct references to parts of the video) and read their press release and some other material about their service. I also spoke with Andras and his co-founder for about 20 minutes at the PMWC; I have not, however, had a chance to watch the hour long video he references below. That said, I have a good grasp of how the backend works here and of what they're trying to accomplish--it's a great idea! I was just unimpressed by the demo video--and startled by the pretty large gap between how they were selling the current iteration of their service and what it actually seems to do.
I'll also note that nothing Andras says below has changed my opinion that I expressed initially that: 1) This has a lot of long-term potential, hence my decision to highlight their service here and 2) Current research into genetics doesn't really support the claims they're making about what their service will be able to do now or in the near-term.
There are a couple places in Andras' note where he implies that I'm somehow out of my element when discussing genetics research. Having researched issues in genetics for several years now, I think Andras and I have a point of disagreement on the state of the field--and I take the condescending tone as, frankly, another example of the shoddy PR work they've been doing--in this note, and in their demo video, which is anything but "sleek."
Finally: I think their push toward broad interoperability and offering users the ability to customize choices based on taste and other factors are both very smart choices. I should have mentioned this in my initial post. They also appear to enable user feedback--an interesting way to collect real-world data--and that could potentially become a very significant source of information and a valuable asset to them. Conceptually, they've done a great job of beginning to think about integrating genetic data with real world information, health conditions, and other sorts of potentially relevant information--so nice job on that front.
There's enormous potential here--just potential that, at the moment, is being wildly oversold.
In any case, I wish them the best of luck.
To Andras' response: