Image via WikipediaIn my opinion the best aspect of the so-called Web 2.0 (a catchy name) is the interaction we get from tools such FriendFeed and Twitter. Accessing these sites, one can be overwhelmed by the amount of information, all kinds of information, ranging from the new photos of tropical fruits, to the end of science as we now it, being liveblogged from ISMB 2008.
In the past months I have been accessing these services/tools more and more, and in the past months by the power of the crowd (at the least the small scientific crowd) a small group of scientists have been exchanging information, ideas, links, tools, comments, learning and teaching anything that comes to mind. There I learned more about the computer cloud (an example here) and one idea leading to another I thought about the possibility of creating an startup that provides bioinformatics computing services on the cloud.
Say you are in a small lab with limited or no access to a large cluster and you have some really interesting data that can analyzed quickly by a cluster, but slowly by a desktop PC, and generate a good set of results for a nice publication. You can always throw the data on a qsub in a university cluster and wait to the big shots to calculate the next prime number before you start your analysis or you can have a personalized virtual image with only the software you need and use the cloud to generate your results and have them when you need. Of course there is a cost to this use, but not close to the cost of applying for a grant to get your cluster, dealing with the sys admin to install the applications you want, dealing with the bureaucracy red tape and finding the right location to put your server.
It has become a cliche to say that (bio)informatics is becoming prevalent in scientific life and research (true but a cliche nonetheless), and it is not difficult to find scientist that don’t have available the infrastructure of a HPC node or a personal cluster. The same way people outsource their sequencing os mass-spec needs, why not outsource their (bio)informatics needs? Many labs have clusters or paid access to clusters even though their use is limited or inconstant. Why invest in a cluster or pay an annual fee while you are using a couple of CPU-hours a week?
When I threw the idea on Twitter and FriendFeed, I didn’t know what to expect. Surprisingly for me there was a good reception to it with some people jumping right into it. I got a nice email from Pawel with some more comments that I think were right on the spot. I don’t see many hurdles to create a startup to provide these kind of services, especially because we live in a virtual world at the moment. There is no need for a physical office, computer infrastructure or employees. The founders, as long as you have a nice and diverse group of people with broad knowledge and interests, can provide all the service and grow together. I am not saying that it will be easy, but after learning a lot from the bright group of people that make our small online community, I think this is a feasible project that can be on the street in a short time.
I am going to prepare a small plan (some ideas on a virtual paper) and anyone interested would be able to join, contribute and start the “company”.