By Katharine Sanderson
A US$2.5-million stimulus grant has been awarded to the American Society for Cell Biology in Bethesda, Maryland, to establish an online open-access database called 'The Cell: An Image Library'. The society is using the money to hire eight part-time annotators, who will help to compile images and videos for the site. Nature caught up with molecular biologist Caroline Kane, from the University of California, Berkeley, who is delaying her retirement to be principal investigator on the project.
Had you always wanted to pursue this project? Or was it just a case of finding a use for stimulus money?
The society started a project like this two years ago, and we had to stop it in December 2008 because of cost. We had only generated about 900 movies and images at the time.
When the National Institute of General Medical Sciences [in Bethesda, Maryland] came up with this grant opportunity, one of the categories was to establish an image library. The institute contacted our executive director and said "we noticed your society was trying to establish a library like this, why don't you please consider applying for this grant".
What's the need for this kind of image bank?
The existing cell image databases (all that we know of) have a limited focus -- on a particular cell type or a particular cell process. You would have to go to that specific database to learn about how your favourite organelle, or cell mutant, looked.
The vision is that the library will be as useful to researchers as the genetic sequence databases are now. That means populating it with cells of all different types as well as images and movies of what's happening inside the cell.
Are the owners of existing databases worried about a super database usurping their work?
Of the people that have some of the larger projects that we've talked to in the United States, the only one that we didn't hear back from in a very positive way we didn't hear back from at all. Everyone else has been very positive about wanting to collaborate, primarily, I think, because it will advance their science too.
So how is it going to help working scientist?
I like to use this analogy: my laboratory works on transcriptional regulation so when I image the cell, I'm interested in what's happening in the nucleus. But the rest of the cell is still there. I work with yeast cells, so if somebody is looking at protein transport or is interested in what happens in a particular mutant series that I have generated, they can look at the structure of the golgi in those cells and see if anything interesting has happened. I don't care about the golgi.
Robert Murphy who's also one of our collaborators, at Carnegie Mellon University in Pittsburgh, Pennsylvania, is developing image-recognition software that if you say 'centriole' can give you a whole list of images of cells where you can visualize the centriole even if the centriole wasn't the primary focus of the person doing the imaging.
How can individual researchers contribute to the databank?
We're going to set up a Flickr account that will enable anybody anywhere to upload images, which will drop into our annotation directory where the annotators can then look at them.
So the annotators will review each entry? That sounds like it's going to take a lot of effort.
Because of my age I'm probably thinking that there will always be human eyes involved. But I think that down the road as the image-recognition software is developed the computer will be able to do some of that initial vetting.
This stimulus grant lasts for just two years. What happens after that?
The word from the National Institutes of Health is that if all goes well then at the end of two years there is the possibility of applying for a contract for five years to continue to develop and expand the database.