Slashdot Mirror
Celera Opens Up DNA Database
greenplato writes "Thirty billion base pairs from the sequences of humans, mice, and rats that were available only by subscription to Celera's DNA database are being put into the public domain. Celera will donate this information to a 'federally run database,' presumably GenBank. Francis Collins, head of the National Human Genome Research Institute, notes that 'data just wants to be public.' Stories in BusinessWeek and The New York Times."
Francis Collins, head of the National Human Genome Research Institute, notes that 'data just wants to be public.'
Data hates when you anthropomorphize it.
They've open sourced me! Does this mean I have to call myself GNU/Steve?
I work for a biotech company with a database which we've been trying to sell subscriptions to for a few years. The prevailing experience with trying to sell the database is that people are very reluctant to shell out the cash to access the data.
I think this is a symptom of trying to sell data to academic institutions. The problems with selling to academic institutions are two-fold; Firstly the universities don't have the cold hard cash to spend on the databases, so any cost over free is too expensive. Secondly, there is the free/open culture within universities that almost punishes commercial ventures for trying to build a business around adding some kind of value to the data (such as convenience or quality of data).
Because of the lack of sales for this database, we're considering handing the data over to a large government body so that they can maintain it, because the company can't simply afford to maintain the database - it costs a lot of money to hire talented people to do database curation.
So when Celera say that "data wants to be free", I think they mean "We'd sell you this data to try and recoup our investment, but we're resigned to the fact that you're not going to buy it".
You can't generally patent "found" sequences.
I wish that were not the case. However, there are many gene patents in existence. The trick is that now you have to show a function for that gene - although bioinformatics is sophisticated (or rather, automated) enough that you can come up with a plausible-sounding function without ever doing benchwork.
What's really being patented is the medical application of these sequences. For instance, Company X discovers that gene Y is overexpressed in cancer Z. They take out a patent on gene Y based on this discovery. That means that no one else can pursue gene Y as a therapeutic target. Moreover, in one case testing for a specific mutation to detect cancer was covered by a patent. This is a very simple piece of labwork being covered, which any competent cancer researcher could have figured out.
The end result is that patents are being awarded for hard work, not for novelty and invention. Throw enough money at a subject, and you'll get data but not necessarily results. Since companies (or academics) can now patent just the data, if someone else gets "lucky" and comes up with an actual result the patent holders can sue the tar out of them if they try to make money off it. (Or even if they don't, as in the case of the breast cancer gene; the company wanted people to pay three times as much for its own testing kit.)
You may soon be able to patent single-nucleotide polymorphisms (SNPs), which may be involved in differential drug responses. Back when I was in college we had a guest lecturer who was a biotech patent attorney, and he said he though SNPs should definitely be patentable. In any case, there is a world of difference between patenting a cancer drug, and patenting a gene (or a FUCKING POINT MUTATION) that may, in the future, be a drug target.
Since most of the human genome is noncoding, I suspect it will be harder to patent pieces of it. I also suspect that some asshole will try anyway.