Dial-the-next-molecule (Liverpool)
One of our Data Science team projects that we have been developing is the molecule recommender which, given an original molecule of interest, will find and display other related molecules and corresponding RSC literature. We investigated, developed, and compared a variety of different ways to define “related” molecules but ultimately were keen to find some “proper chemists” to show these results to, record what were most useful, and not useful, and what they might use such a tool for.
The back-to-back “Dial-a-Molecule Annual Meeting - Transforming Chemical Synthesis” and “RSC CICAG group meeting - Chemical Structure Representation: What Would Dalton Do Now?” meetings were an excellent opportunity to do just that - meeting old friends and colleagues that we have worked with in the past and making new contacts for the future.
At the former meeting we gave a flash presentation to encourage people to come over to the Royal Society of Chemistry stand to try out the Molecule Recommender (or sign up to do so at a later date) as well as find out more about our journals and membership services. At the latter meeting our aim of putting molecules in context with their related publication data and related molecules, beyond their immediate drawn structure reflected the wider theme of the meeting and we had an opportunity to give a fuller description of the approaches taken and took part in the final panel discussion.
As well as being useful to gain immediate direction for our efforts, both meetings set out visions of organic synthesis and chemistry as a whole (put in the context of the rich history of chemistry) which featured robots, artificial intelligence and potential machine overlords. If you were unable to attend and would like to be inspired, the slides are available one the CICAG event webpage. These include Aileen Day’s presentation about the molecule recommender and particularly philosophical overviews from our previous collaborators Jeremy Frey (Southampton University), Roger Sayle (NextMove) and Jonathan Goodman (Cambridge University).
If you would like to try out the molecule recommender and let us know what you think, please contact us.
Contact Data Science
- Email:
- Send us an email