A few weeks ago, I did some "back-of-the-envelope" calculations to explain to a reader why genome sequencing costs so much.
I estimated that, if JCV's genome were sequenced at the cost advertised by university core laboratories, his genome would cost about $128 million.
That was an estimate, of course. But what did it really cost?
Genome Technology asked J. Craig himself. In the October 2007 issue of GT, JCV estimates that the cost from the first Celera human genome project (guess who?) was about $100 million and that the cost of his most recent genome project was at least $70 million.
Interestingly, J. Craig also said that his will probably be the last human genome to be sequenced by capillary electrophoresis and Sanger dideoxy sequencing because of the cost.
Next generation sequencing, here we come!
That's the same figure Venter gave me in an interview that will be posted within a few days at bloggingheads.tv The thousand-dollar genome is a ways off.
The thousand dollar genome is a long ways off if we try to obtain it via capillary electrophoresis and Sanger dideoxy sequencing.
The next generation sequencing technologies are an entirely different story. If we can work out the data management and analysis challenges, we'll certainly have thousand dollar genomes within the next decade.
In recent items about further sequencing at the Venter institute, the estimate for sequencing with the ABI SoLID platform was in $300K/genome -- still a ways from the $1K genome.
Without independent auditing, all of these numbers have serious error bars -- and depend on all sorts of locality-dependent figures & accounting fudge factors.
Also, it is interesting to see that the recently announced 'Chinese genome' may have been done on Solexa instruments -- until a real paper comes out, the newswire sources require a lot of interpolation to figure out what was done.
I'm sure to be learning quite a bit more about next generation sequencing and the costs after Thursday. We're pretty focused on it these days.
One reason the next-gen technologies are even being discussed for sequencing human genomes is because there is a quality assembly upon which the next-gen sequences can be assembled. While we may be a few short steps away from next gen sequencing for resequencing human genomes, I wouldn't get my hopes up for de novo next gen whole genome sequencing projects of many eukaryotes. Of course, for those people interested only in human genomes (ie, most people), the inability to use next-gen sequencing for de novo genome projects is a minor issue.
Watson's genome was a de novo project, carried out with 454 sequencing.
I think what RPM is saying when he means de novo is organisms that have no reference assembly available, so they have anchor to align reads to.
I am getting to play around with Solexa data, and data management is an issue! I couldn't even view my data with Consed because my machine didn't have enough memory!
I agree, I would define "de novo" sequencing as sequencing a new sample without a reference sequence, as well.
I thought, though, for some reason that 454 claimed that they didn't use alignments to a reference and that it was de novo.
I'll have to check on this though to find out for certain.