I’m tired of trying to pretend that I follow college basketball, and looking at an empty bracket fills me with anxiety. What’s a sports-adjacent person to do? Well, if you’re like me, you simulate the entire tournament using win probabilities from 538 and let the pseudorandom whims of fate decide your bracket’s destiny. If you want to deploy RoboBracket3000 in your own office pool, it’s a short python3 script available on GitHub: https://github.com/jrvalcourt/marchmadness
Category: Fun
We imagined a satirical genomics startup and accidentally reinvented Vinome
Two days ago, my roommates and I were trying to concoct a satirical genomics startup. We wanted an idea so ridiculous that no one could possibly believe it was real. Instead, we accidentally reinvented Vinome.
Vinome claims to be “Your DNA Guide to Wines You’ll Love,” a service that will, for a cool $110 fee (or $30 if you’ve already used another Helix application), predict which types of wine you’d enjoy imbibing. Oh, and as a convenience, they’d be happy to sell you those wines. It’s “taste proven by evidence,” without a shred of evidence.
Part of Vinome’s claims seems to rest on work that links some genetic variations with a person’s sensitivity to certain tastes. For example, a few genetic regions contain a bit of information about how sensitively one can perceive bitterness. But knowing whether a person is more likely than average to be able to taste certain elements of cilantro is a far cry from knowing anything about what kind of wine they’d like. The idea that such a complex personal preference could be meaningfully determined by one’s genetic makeup is absurd. A rule of thumb: if your fun genomics company relies on being able to do something that most scientists believe to be impossible, you should either step up and claim your Nature paper or stop misleading people.
(Vinome also claims to have conducted a study that helped them define their “algorithm,” but it hasn’t been published yet and the details are sketchy. To put it gently: I’ll believe it when I see it.)
Companies that claim to read information from the genome that simply isn’t there abound—see the infamous Soccer Genomics, an idea so farcical I’m still not entirely convinced it’s meant in earnest—but Vinome is more troubling. As a part of the curated Helix ecosystem of DNA tests, Vinome has supposedly passed some level of scrutiny. Helix, which was founded in part by the sequencing company Illumina, aims to be a one-stop-shop for DNA tests of all kinds, but Vinome’s presence on the platform is cause for concern. If Vinome earns Helix’s (and, by extension, Illumina’s) imprimatur, it can’t be good news for the rigor of this new marketplace.
p.s. You can find that satirical startup we imagined before I found out Vinome was a thing here.
Update 10/13/17 @ 3 PM: I’m chatting with Vinome over on Twitter. Once I have a sense of their position, I’ll update this post.
Update 10/26/17: Vinome hasn’t responded to most of my questions, but they have reiterated that their study is still in preparation.