It is time for a whole new approach. Front-line biology cannot function without these resources, so solutions must be found at both national and international levels.
That’s from an editorial in Nature talking about the need to change the way data resources are funded, a topic I have visitedbefore.
In this particular case, the resource under question is The Arabidopsis Information Resource (TAIR), which had been funded by the NSF for a decade but is seeing that sole source of funding phased out. As the editorial points out, such challenges are not unique to TAIR, but a problem facing biological databases and repositories, i.e. part of the core infrastructure for the life sciences. I say that because to sustain any data resource you need to sustain infrastructure as well as the development work required to keep the data resource viable and functional.
What is the right model? It’s a combination of factors. We needs a realization that biological databases are core to the science of today, and need to be sustained and not have to depend on short term grants; an understanding that you need to invest in software development; that it might be better to encourage scientists to contribute and leverage existing resources and that not every database is worth funding. It could involve a mix of philanthropic contributions, government funding and even private companies who seek to and want to use public data sources. Some have noted that being able to mirror data sources, etc is important, and in some cases that makes a lot f sense. After all it’s easier to fund a smaller data source and perhaps there could be meta-repositories that use APIs to pull in data when required. But one thing is clear from experience. Sustaining a data resource of any kind requires a dedicated team, a team of software engineers, curators, informaticians and biologists and funding that understands that these need to be maintained. Of course, I also believe that you can do this without capital investment today, but that’s a detail. What we need to do is realize that our current model just doesn’t work for biological data resources
Funding biological data resources revisited
That’s from an editorial in Nature talking about the need to change the way data resources are funded, a topic I have visited before.
In this particular case, the resource under question is The Arabidopsis Information Resource (TAIR), which had been funded by the NSF for a decade but is seeing that sole source of funding phased out. As the editorial points out, such challenges are not unique to TAIR, but a problem facing biological databases and repositories, i.e. part of the core infrastructure for the life sciences. I say that because to sustain any data resource you need to sustain infrastructure as well as the development work required to keep the data resource viable and functional.
What is the right model? It’s a combination of factors. We needs a realization that biological databases are core to the science of today, and need to be sustained and not have to depend on short term grants; an understanding that you need to invest in software development; that it might be better to encourage scientists to contribute and leverage existing resources and that not every database is worth funding. It could involve a mix of philanthropic contributions, government funding and even private companies who seek to and want to use public data sources. Some have noted that being able to mirror data sources, etc is important, and in some cases that makes a lot f sense. After all it’s easier to fund a smaller data source and perhaps there could be meta-repositories that use APIs to pull in data when required. But one thing is clear from experience. Sustaining a data resource of any kind requires a dedicated team, a team of software engineers, curators, informaticians and biologists and funding that understands that these need to be maintained. Of course, I also believe that you can do this without capital investment today, but that’s a detail. What we need to do is realize that our current model just doesn’t work for biological data resources
Related articles by Zemanta