We’re pleased to highlight another great plugin in our Spotlight series of posts. This week, we’ll be profiling STITCH. Without further ado:

In one tweet (140 characters) or less, introduce us to your website.
STITCH is a database of interactions between proteins and chemicals. It can be used for interactive exploration and for large-scale studies.

Why is your database unique and special?
As far as we know, STITCH is the only publicly available database that combines data on protein–chemical associations from experimental and manually curated databases and text-mining into a comprehensive overview on what is know on a particular small molecule. Imported sources include: BindingDB, PDSP Ki Database, GLIDA, DrugBank, CTD, PharmGKB, KEGG, and PDB. However, STITCH does not make any predictions for chemicals that have not been deposited in our source databases and have not been mentioned in the literature.

Who is your target audience?
STITCH is useful for everyone who wants to find out what is known about a particular chemical. Thus, our target audience is scientists working on molecular biology and pharmacology. For example, I’ve used STITCH to check a list of compounds that showed activity in a high-throughput cellular assay to see if anything is already known that might link the compounds to the observed activity.
Computational biologists can use it as a dataset for their studies. The research group of Andreas Bender used the STITCH network to automatically extract the targets of anti-tuberculosis compounds in Mycobacterium tuberculosis (Prathipati et al.). We ourselves used the data on which STITCH is based to build a drug–target network to predict drug targets using side-effect similarity (Campillos et al.).

What improvements are coming in the future?
Of course, we will add new sources to STITCH and update the sources we already imported. The main planned improvement is to provide the option to look at chemicals at different resolutions: Currently different salt forms and stereoisomers are mapped onto one record. Ideally, the user should be able to choose if they care about stereochemistry or if they’d like to have all information combined (as some data sources might not identify the relevant isomer).

Thanks to Michael Kuhn for helping us understand STITCH more, and be sure to check out the STITCH plugin in BioGPS.