Archive for the ‘pca’ tag
My previous post did a quick comparison of the GSK anti-malarial screening dataset with a virtual library of Ugi products. That comparison was based on the PubChem fingerprints and indicated a broad degree of overlap. I was also interested in looking at the overlap in other feature spaces. The simplest way to do this is to evaluate a set of descriptors and then perform a principal components analysis. We can then plot the first two principal components to get an idea of the distribution of the compounds in the defined space.
I evaluated a number of descriptors using the CDK. In a physicochemical space represented by the number of rotatable bonds, molecular weight and XlogP values, a plot of the first two principal components looks as shown on the right. Given the large number of points, the plot is more of a blob, but does highlight the fact that there is a good degree of overlap between the two datasets. On going to a BCUT space on the left, we get a different picture, stressing the greater diversity of the GSK dataset. Of course, these are arbitary descriptor spaces and not necessarily meaningful. One would probably choose a descriptor space based on the problem at hand (and also the CDK XlogP implementation probably needs some work).
I was also interested in the promiscuity of the compounds in the GSK dataset. Promiscuity is the phenomenon where a molecule shows activity in multiple assays. Promiscuous activity could be indicate that the compound is truly active in all or most of the assays (i.e., hitting multiple distinct targets), but could also indicate that the activity is artifactual (such as if it were an aggregator or flourescent compound).
This analysis is performed by looking for those GSK molecules that are in the NCGC collection (272 exact matches) and checking to see how many NCGC assays they are tested in and whether they were active or not. Rather than look at all assays in the NCGC collection, I consider a subset of approximately 1300 assays curated by a colleague. Ideally, a compound will be active in only one (or a few) of the assays it is tested in.
For simplicities sake, I just plot the number of assays a compound is tested in versus the number of them that it is active in. The plot is colored by the activity (pXC50 value in the GSK SD file) so that more potent molecules are lighter. While the bulk of these molecules do not show significant promiscuous activity, a few of them do lie at the upper range. I’ve annotated four and their structures are shown below. Compound 530674 appears to be quite promiscuous given that it is active in 46 out of 84 assays it’s been tested in at the NCGC. On the other hand, 22942 is tested in 232 assays but is activity in 78 of them. This could be considered a low ratio, and isoquinolines have been noted to be non-promiscuous. (Both of these target kinases as noted in Gamo et al).
I recently described a REST based service for performing PCA-based visualization of chemical spaces. By visiting a URL of the form
one would get a HTML, plain text or JSON page containing the first two principal components for the molecules specified. With this data one can generate a simple 2D plot of the distributions of molecules in the “default” chemical space.
However, as Andrew Lang pointed out on FriendFeed, one could use SecondLife to look at 3D versions of the PCA results. So I updatesd the service to allow one to specify the number of components in the URL. The above form of the service will still work – you get the first two components by default.
To specify more components use an URL of the form
where mol1, mol2, mol3 etc should be valid SMILES strings. The above URL will return the first three PC’s. To get just the first PC, replace the 3 with 1 and so on. If more components are requested than available, all components are returned.
Currently, the only available space is the “default” space which is 4-dimensional, so you can get a maximum of four components. In general, visit the URL
to obtain a list of currently available chemical spaces, their names and dimensionality.
While it’s easy to get all the components and visualize them, it doesn’t always make sense to do so. In general, one should consider those initial principal components that explain a significant portion of the variance (see Kaisers criterion). The service currently doesn’t provide the eigenvalues, so it’s not really possible to decide whether to go to 3, 4 or more components. For most cases, just looking at the first two principal components will sufficient – especially given the currently available chemical space.
Update (Jan 13, 2009)
Since the descriptor service now requires that Base64 encoded SMILES, the example usage URL is now invalid. Instead, the SMILES should be replaced by their encoded versions. In other words the first URL above becomes
http://rguha.ath.cx/~rguha/cicc/rest/chemspace/default/ YzFjY2NjYzE=,YzFjY2NjYzFDQw==,YzFjY2NjYzFDQ0M=, Qyg9TylDKD1PKQ==,Q0MoPU8pTw==
The ONSChallenge has been running for some time now and the simple web query form that tied in the data from Google Docs along with web services from IU has turned out to be pretty handy. With more and more data becoming available, I had done some initial exploratory analysis of the measured solubilities. One thing that is useful to the experimentalists is a suggestion of which compound to test next. This could be made on the basis of many factors – availability, ease of synthesis and so on. But one way to look at it is to examine what types of compounds have been tested previously, and suggest that the subsequent compounds be very different from those that have been tested.