Sixty-Four Free Chemistry Databases Part 13: Exploring Drug Side Effects with SIDER
Chemistry databases have a way of building on top of one another to reveal new insights. In today's installment of our continuing series on free chemistry databases and Web services, we take a look at Side Effect Resource (SIDER).
SIDER is organized around the side-effects observed with marketed drugs, as determined primarily by product inserts. Relationships between drugs and side effects can be explored by browsing by the names of side effect or drugs. Text-based searching for either side effects or drug names is also possible.
Finding connections between side effects and drugs is especially easy due to the way these relationships are hyperlinked. For example, the entry for Gabapentin lists all known side effects of the drug, each of which are hyperlinked to a page dedicated to that side effect. Clicking on Ataxia, for example, I'm taken to a page that shows me that this side effect is shared by dozens of other drugs, one of which is Risperidone. Going back to Gabapentin, we can click within the colored data presentation to find the exact primary reference being cited. Simply beautiful.
A very interesting feature of SIDER is the visual presentation of side effect profiles. Again referring to the entry for Gabapentin, we see a color-coded map of the side effect profile that concisely summarizes a wealth of human clinical data.
SIDER has many applications, one of the most interesting of which is the prediction of drug-target correlations for non-analogous chemical structures. For those with an interest in data mining, SIDER's underlying dataset is freely downloadable under various forms of Creative Commons licenses.
One of the most interesting things about SIDER is how it was created: completely from information available to the public. One of those sources is the STITCH database, which maps interactions of small molecules and proteins. It should come as no surprise that STITCH itself uses PubChem. In other words, SIDER relies indirectly on PubChem for mapping drug label information onto chemical structure.
For those looking for information on side effects of marketed drugs, or ways to use that phenotypic information to discover new relationships, SIDER is an excellent resource.
Kudos
- Streamlined, elegant design.
- Color-coded information summary.
- Primary sources included for some data.
- Data freely downloadable.
- Global statistics and search interface on home page.
Ideas for Improvement
- More prominent display of chemical structures - side effect summary pages like this one are a good example of where this would be very helpful to medicinal chemists.
- Hyperlink to source of primary information.
- Add the name of side effects and drugs to URLs for better search engine discoverability.
- Feeds for new drug or side-effect entries.
I've looked through SIDER. It's a great interface and clear how to use and interpret the data. Unfortunately the database suffers the inadequacies of deriving the structure based data via STITCH via PubChem. For example, a number of the structures are simply not what they are labeled to be. An example is Taxol which has many stereocenters but on SIDER is lacking all stereochemistry: http://sideeffects.embl.de/drugs/4666/all . I noticed the same issue on Wolfram. it is up to the user as to whether this is important or not of course.
Hi,
We have just added your latest post "Sixty-Four Free Chemistry Databases Part 13: Exploring Drug Side Effects with SIDER" to our Directory of Science . You can check the inclusion of the post here . We are delighted to invite you to submit all your future posts to the directory and get a huge base of visitors to your website.
Warm Regards
Scienz.info Team
http://www.scienz.info
@Tony - good point. I wouldn't quite put lack of stereocenters into the category of not being "what they are labeled to be.' I'd say instead that the structure is low-resolution. As far as the taxol depiction goes, it appears to be accurate. It just doesn't go down to the level of resolution that other sources might.
If other structures on the site do show stereochemistry, that's more of a problem, because it implies that the stereochemistry of taxol is unknown. This doesn't appear to be the case, however.
I'm curious, what steps would you suggest to the creators of SIDER to improve the quality of their chemical structures?
I followed the link: [side effect summary pages like this one] cited in the original post and I clicked on the pregabalin link. It showed a structure with stereogenic center marked. It seems to be derived from PubChem. I followed the link to Wikipedia and it gave the enantiomer for this substance. ChemSpider and OpenEye agree with Wikipedia - probably no surprise.
Stereochemistry affects adverse effects compare the enantiomers of penicillamine or the enantiomers of thalidomide if you need confirmation. If structures are reported and stereoisomerism is important to the activity, then it must be reported and correct.