SciBite - Open Intelligence on Pharmaceutical Discovery & Development
Lee Harland, is a visitor in the ChEMBL group here at EMBL-EBI, and he is collaborating with us on semantic web data integration, text-mining, target ontologies and so forth. I asked him to write a small piece for the ChEMBL-og on one of his personal projects - SciBites, and here is what he wrote......
"SciBite is a new biomedical alerting service tailored to pharmaceutically relevant questions and focused on targets, diseases and drugs. The premise is simple, right now if you're a scientist interested in say, Asthma, how can you stay on top of all the lastest developments? You can of course, set up pubmed searches, patent searches, google news searches (making sure you have remembered every possible synonym as these tools won't do that for you). What you'll get back is a stream of articles. Some relevant, many not, and you'll have to read them to actually find out. We thought there had to be a better way.. What we really wanted was a kind of Twitter for "things", where the drugs, disease and targets could each tweet relevant news, in an easily consumable (and discoverable) way. So we built SciBite.
SciBite works by continually scanning 1000s of sources including literature, patents, news feeds, blogs, databases and more. It looks at every new article and automatically tags the targets, diseases, drugs, companies and 'contexts' (such as "biomaker study", "regulatory approval", "animal model" etc), that it finds. Users can then go to the website and view information by topic, not by source. The lists can be filtered based on source or other criteria such as regulatory approval or biomarkers. Everything is available as an RSS feed so users can stay on top of latest news. Information is presented in a very visual way, making it really easy to scan new articles and identify the key topics. We also have relevancy filters to remove spurious and irrelevant news (although this will never be perfect!). Finally, our "related topics" function allows you to quickly find the targets, diseases or drugs being co-mentioned with the thing you're interested in, which is an incredibly powerful way to spot new and interesting connections.
There are a number of major content companies doing this sort of thing, and as a small (tiny) company, we cannot hope to match their levels of curation and resources. However, we believe that this sort of information should be available to everyone and so we hope that by using technology we can provide something that comes close to what the major players offer. We decided early on that as we were using a lot of public data (including ChEMBL), the site itself should be free and support these efforts. We're also making all our data and APIs freely available to any non-profit organisation expressing an interest.
We've seen a great growth in user numbers since our very low-key launch at the start of February. Over 2000 people have used the site, and we've done hardly any advertising... Its daunting to see that many so soon, but its a great feeling to know people are finding it useful! What's there now is really just the start. The aim was to build a platform that connected news to things, which we've done. The next stages are to do much more with the data. This is one of the things I'll be exploring as part of my Visitorship with the ChEMBL group, we're looking at some interesting company-centric profiling, tracking whats going on with each organisations drugs. There's a whole lot more planned for 2012 too!
Anyone interested can use the system for free now, at http://scibite.com. I tweet as @SciBitely and our blog is at http://about.scibite.com."
Wow! Look at what is there.