The power of hype

First, a quick nod to Nigeria which has begun its relatively successful elections this week. Here’s hoping that the term “relatively” will be dropped over time. Also, I’m keeping an eye on developments in Yemen, to see which way the army breaks as the GCC seeks to peacefully remove Saleh. As I mentioned previously, it looks like elements of the army are lining up for reform so we may see more-or-less peaceable transition. Finally, it appears Gbagbo is captured and the democratically elected leader of the Ivory Coast can start governing fully. And all without a single NATO no-fly zone.

But that’s not what I wanted to write about today. Instead, I’d like to call your attention to Steven Pinker’s My Genome, my self article from the New York Times Magazine. I came across it in the book I’m currently reading: The Best American Science Writing of 2010. The article is by no means succinct but it is an immensely readable detailing of the current state of human genomic research.

He lays out the potential and limits of current understandings as well as the hype surrounding the “personal genomics” movement over the past decade. As a genomics researcher these past few years, I came in right when grand ideas were becoming actual projects and have watched as they stumbled across the always-unexpected hurdles. It is a pleasure to read Pinker’s layman’s guide to narrow band that is actually revealed from current academic and commercial testing.

Basically, the big labs and companies can tell you with decent accuracy whether you have certain known, popular mutations. Of course, you may realize this means they can’t reveal whether you have less common mutations or even unique ones! The companies will also quickly let you know that this mutation (really a single nucleotide polymorphism or SNP) is associated with this disease/trait/ethnic group. But what they fail to mention is that the strength of that association is almost always extremely weak.
Pinker’s example is height. Long discussed at academic conferences, height is the perfect trait: quantitative and highly heritable. Looking for common SNPs in tall or short people has revealed a host of associations. But when you look at all the known ones, they account for only 2% of the heritability! This means that while it’s almost automatic that tall couples make tall kids, tall kids don’t always inherit the known “tall” mutations from their parents. Since it’s easier to find SNPs with strong effects (think Huntingdon’s disease, where the SNP explains 100% of the heritability), and we’ve found dozens of SNPs associated to height, that means there are at least hundreds other SNPs that are contributing. So maybe tall kids inherit 50 from one parent and 50 from another, but we only are looking at 2.

What’s worse is that different studies don’t always find the same SNPs. Studies looking for associations in thousands of papers often come up with different, non-overlapping lists of SNPs. So which ones are right? Well, right now, we just say they’re all correct. But of course the studies are done in different ethnicities, and given that genes are highly interactive, what may be height-enhancing in Scandinavians is meaningless in Japanese.

We “finished” the human genome project a decade ago but really are not much closer to that heralded age of personalized medicine. It’s my opinion that we’re never going to get to that predicted moment where you submit a DNA sample and get an ideal drug-cocktail and health-warning. There’s simply too much genetic individuality that will never be understandable. Sure, we’ll be able to sequence genomes for pennies soon, but if you’re the only person in the history of the world with a set of specific mutations who can possibly predict whether that makes you more or less susceptible to statins or HIV or skin cancer or poor inhibition control.

That’s not to say this research is useless. It would be better applied to things like agriculture and biotech. Herds of cattle and rows of corn have less genetic differences so we can make conclusive findings that tie SNPs with trait differences. Indeed companies like Monsanto are gobbling up the biotechs that make just these finds. These natural polymorphisms should calm those afraid of genetically-modified crops as they can just be bred into the next generation and are entirely native.

Media type: Book
From: The Best American Science Writing 2010
Title: My Genome, My Self
Read it at: