Saturday, July 23, 2011

The $1,000 genome: Coming to you in October 2012!

The National Human Genome Research Institute (NHGRI) of the National Institutes of Health (NIH) provided funding and leadership for the United States’ participation in the Human Genome Project. Since the first human genone was sequenced, the NHGRI has continued supporting the development of new applications of genomics, including human re-sequencing, the sequencing model organisms, developing new technologies, and creating new data analysis methods. In short, they have their finger on the pulse of genomics.

A few months ago, I stumbled onto a page on the NHGRI website page that tracks the cost of sequencing since 2001—shortly after the first human genome was completed. What’s clear in their “cost per genome” graph is the steady decrease in cost that has occurred over the past decade.
Looking more closely at the graph, you’ll also notice a dramatic change in its slope starting in late 2007—a striking shift that coincides with the introduction of “Next Generation” sequencing technologies from Applied Biosystems and Illumina. Since then, the cost of sequencing has been falling at a truly astounding rate.
If you download the Excel spreadsheet containing the data NHGRI used to generate its cost graph, and do a little analysis, you’ll find that the cost of sequencing has been falling by one third each quarter. Ignoring slight fluctuations, this drop has been as close to consistent as one could imagine, which allows us to extrapolate where costs are headed.

Although NHGRI makes some very “old school” assumptions about sequencing, including the use of a mix of new and old technologies, it ignores recent announcements, such as Illumina’s $4,000 genome, Knome’s $4,998 genome and analysis package, and BGI’s “summer special” $999 exome sequencing. Consequently, the NHGRI estimate is much higher than the “street price”: for sequencing; when  Illumina made its $4k announcement in April, the NHGRI sequencing cost estimate was $13,835.

But let’s be conservative and start from NHGRI’s estimate. If we assume a one-third drop every quarter, the cost of sequencing will fall to $1,143 by October 2012. Extrapolate further, and the cost will drop to $94 by April 2014. (See graph below.). And remember, this ignores current prices as well as the potential for third generation technologies. If PacBio, Oxford Nanopore, NabSys, GnuBio, Ion Torrent, or one of a dozen other players comes out with a new machine, who knows how fast costs wil fall?

Regardless, the big question is no longer when the $1,000 genome (or even the $100 genome) will arrive but rather, what we’ll do with the data once we have it. It’s time to start thinking about the scenarios that will present themselves everywhere, from doctor’s offices to high-school science labs, once genome sequencing becomes less expensive than two tanks of gasoline or a month’s worth of lattes at Starbucks.


Tuesday, June 28, 2011

The future of genomics is (almost) now

In June 2000, Francis Collins and Craig Venter stood with President Clinton and announced that the first draft sequence of the human genome had been completed. At the time, there was a great deal of excitement about the potential of this achievement and how the genome sequence promised to forever change medicine. Yet, eleven years later at my annual physical, my physician didn’t pull out a genome tricorder to scan my genome. Instead, he put on a glove and probed and prodded me using examination methods that date back to Hippocrates himself.

Having been involved in genome science since its early days, I didn’t find this surprising. It had always been clear to me that the draft genome was just a reference and that we would need data from a large number of people before we could begin to make sense of what was encoded in the genome. Back then, however, the cost of sequencing was too great to seriously consider sequencing the genomes of hundreds, or thousands, of people.

Sequencing the first genome cost about of $3 billion and took almost 15 years to complete. New technologies that were developed along the way dramatically decreased the cost, reducing the time it took; but it was not nearly enough. In January 2010, my research sequenced an ovarian cancer genome and a germline genome from the same person—a project that took three months to complete. But in the past eighteen months, things have changed so dramatically that we’ve now begun to think that we will be able to sequence (almost) everyone.

In just the last few weeks a number of events suggest this day is close at hand. First, Illumina announced they would sequence an individual genome for $5,000—and discount the price to $4,000 for 50 or more. This is about the same cost offered by Complete Genomics. And today I received an e-mail from Knome offering genome sequencing and data analysis for $4,998. Not to be outdone, the Beijing Genome Institute e-mailed me its offer to sequence the exome—the known portion of the genome that encodes proteins—and analyze the resulting data for $999.

Given how costs are falling, I can easily predict that sequencing a human genome will cost about $2,500 within a year, fall to $1,000 within two years, and be $100 or less within five years (likely sooner). This changes everything.

I used to say, in all seriousness, that if I developed cancer I would mortgage my house to have my genome sequenced. Today, I could probably pay for the service with my credit card. But at a price of $1,000, I would have my wife’s genome sequenced for her birthday. And for $100, let’s face it, we would have both our genomes sequenced instead of going out for a Michelin-star meal. And I don’t think we’d be alone.

It’s simply no longer a matter of if but of when we’ll start collecting data on ourselves—and on everyone else. The questions now are what are we going to do with these data once we have collected them—and who will have rights to use those data and in what way?