In Nature this week (sorry kids - subscriptions only), we find both a correspondence article and a news article, the latter entitled "Databases in peril". The gist of these stories is that lack of long-term funding may threaten the future of public biological databases.
This got me thinking about database maintenance in general. Not having a business mind, I have never given much thought to the costs of maintaining resources such as databases. The news article claims that the Mouse Genome database costs USD 4 million per year to run. That seems like a lot to me, but when I consider the modest web resources that I maintain, I suppose there are components from:
- my salary
- hardware repair and upgrade
- electricity and bandwidth costs
just off the top of my head. Not to mention the problem of who maintains these resources should I leave/run away/die.
Have we created resources that could and should last "for ever"? Can they?
Comments
distribute it...
Maybe distributed databases could help dilute the costs, using grid or p2p architectures.
-------------------------------
Marcos Oliveira de Carvalho
Southern Genome Investigation Program
http://www.cbiot.ufrgs.br/~wwwgenoma
Nothing lasts for ever !
I work for the EBI and the nature article is true in many respects. The EBI is one of the three biggies in the biological database field. However, the EBI does not have a source of fixed funding in the long term. So, in theory, the institute could run out of money someday and all the databases that are created, maintained or mirrored here could just vanish !.
As money comes in from the EU and various other sources, they are for fixed terms ranging from a couple of years to about 5 years. At the end of each such term, new grants have to be applied for, fresh cases made for support, and new sources of funding researched. Of could always argue that such long term funding may lead to complacence and laxity, but the truth is that uncertain funding more often leads to hastily cobbled databases in order to fulfil certain 'deliverables' in the grants.
Take the example of HGMP, a website for providing resources for the Human Genome is to be shut down in another month since the MRC, UK made a decision to close down the bioinformatics divisions just a year after the institute was renamed as the Rosalind Franklin Center for Genomics Research. This appears to be have been as a result of MRC wanting to cut down on the funding for bioinformatics research and use the money for the upkeep of hospitals etc.
In my opinion, resources such as the NCBI, EBI or DDBj are really important to the scientific community and the resources they offer should stay.
Measure of the usefulness
I agree with the previous post. I would just add that there should be some measure of the usefulness of the database to the community in general before it is granted a fixed maintenance funding. There are some databases like pubmed, pdb, sequence databases, etc that are so commonly used and are so fundamental for everyday science that they should not depend on development grants.
agreed
I concur. Research without the NCBI seems unthinkable.
Then again - none of this existed 25 years ago. I wonder how fundamental biological data will be stored and accessed in 25 years time? Perhaps in a way that we can't even imagine at the moment, perhaps much as now but "bigger, faster, cheaper".