Basically, we're currently at a stage where costs are dropping much faster than the rate at which the value of genetic information is increasing - in other words, you and I will be able to afford a genome sequence long before science will be able to tell us much about what it really means. However, the beauty of a genome sequence is that (unlike 23andMe-style chips, which are constantly being replaced by higher-resolution models) it never becomes obsolete. Once you have your entire sequence sitting on your hard drive you can simply sit back and wait for new associations and techniques for assigning function, which will soon be appearing at an exponentially increasing rate as technology improves.
In other words, the data appreciates. But unlike buying a home mortgage, there is going to be a mass drop in the cost in the near future. This figure illustrates the point effectively. To some extent then the analogy to the housing market breaks down. In many regions of the world fixed stock of land means that values will increase over the long term (since populations and per capita purchasing power are increasing) so "getting in early" makes sense. Rather, it's like buying a Kindle; everyone is going in that direction at some point but early adopters are going to pay a premium for gaining access to that utility bundle before the rest of us. Of course, the very fact that we're mooting these choices says something about the state of the game.
I disagree with your statement that genome sequencing is at its "final stage" - that if you get a full genome from current sequencers you have everything youÂ´ll ever need.
First, current sequencers produce strands that are too short to acquire haplotypic/phase information, which will certainly be a very significant factor in gene expression, and therefore phenotype. Second, there are numerous other epigenetic modifications such as methylation and X-inactivation which will also affect expression. Third, mitochondrial DNA will possibly also influence common diseases and should be analyzed.
Current sequencing technology "rocks", but its just the begining...
where'd i say "final stage."?
I guess you're referring to my statement rather than razib's?
I should have clarified this in my post. Basically, I agree that current technology can't assemble a complete genome sequence in a meaningful sense of the word (there's a disclaimer at the bottom of my post to that effect). However, long-read technology is very close; there's little doubt we'll see a commercial platform with long-read single molecule sequencing capability within a few years (if not from PacBio then someone else - the market is starting to look pretty crowded).
Once we have a sequencing platform that can accurately scan 10,000 or so bases from a single molecule in an ultra-high-throughput fashion, we can build a genome sequence that is as close to "complete" as you'd need it to be for most practical purposes (i.e. missing only a few nasty, large repetitive regions). That's the sort of sequence I envision being able to put up on the shelf and take down every few weeks when new association results roll in.
Mitochondrial DNA will be sequenced as a matter of course by most platforms (e.g. the data currently rolling out of the 1000 Genomes Project includes massive coverage of the mitochondrial genome, because it's present in so many more copies than the nuclear genome).
Epigenomics is a wild-card - I think most people accept that epigenetic modifications are likely to be important, but it's totally unclear how important they will be. Now that is something that you can't just do once and then forget about it; epigenetic modifications vary by tissue, age, diet, even time of day. Epigenomic monitoring might end up being one of those things we get done every few years, like a blood test - who knows?
Since the axes of Cost and Information can be scaled independently, why not move the "now" line to just after the intersection of the two lines?
That would better express the movement, albeit small, of the use of genetic sequencing from the research lab to the market place and clinic.
The area represented by information minus cost should approximate yield: when cost is above information, yield is negative and represents a capital infusion. When information is above cost that represents return on investment. If the industry stopped investing today, they could make a small profit; we just don't see that because they are scrambling to pour venture money into their businesses for a gargantuan profit in the future.
This chart also does not assume boundless exponential growth in information; I think it should. Existing variation is immense but finite, which supports the bounded curve. Novel/artificial genes are not finite, thus the value of the genetic enterprise has no limit.
E.g. Genes that can give new powers or extend lifespan.
In a similar vein, the cost curve originated not at some high, but theoretically attainable plateau. It came out of the asymptotic realm of impossibility and has recently been roughly 1/1000th the price each subsequent decade.
Will it reach 350$ in 2019? I think that is very doable, if "4th gen" is something like the microfluidics being done at Duke and elsewhere. http://microfluidics.ee.duke.edu/ My epithelial cells duplicate and "read" my genome for less than a penny per day!
Thanks for the blog. I read it everyday.