I'm please to announce that my PhD thesis is now available from the Cambridge DSpace repository: https://www.repository.cam.ac.uk/handle/1810/246652. One thing potentially of note is the description of fast Kekulisation that I originally intended to write as a blog post. Also following up from NextMove Software's recent post by Daniel on Cahn-Ingold-Prelog (CIP), the results of Chapter 6 contains some more CIP madness.
Tuesday, 30 December 2014
CDK 1.5.10 has been released and is available from sourceforge (download here) and the Maven central repository (XML 1).
This release follows very shortly after 1.5.9 and is the first release available from the central maven repository. This means there is now no need to include a custom repo when using the library in downstream projects (XML 1)
The short release notes (1.5.10-Release-Notes) summarise and detail the changes. Other than the availability in the central repository the release includes a new MolecularFormulaGenerator contributed by Tomáš Pluskal that provide mass to formula generation in a fraction of the time of the old MassToFormulaTool.
<dependency> <groupId>org.openscience.cdk</groupId> <artifactId>cdk-bundle</artifactId> <version>1.5.10</version> </dependency>
Wednesday, 24 December 2014
CDK 1.5.9 has been released and is available from sourceforge (download here) and the EBI maven repo (XML 1).
This is the first release to be built using Java 7 and will require the Java SE Runtime 7 to execute. The previous release (1.5.8) will be the last to work with Java SE 6.
The full release notes (1.5.9-Release-Notes) summarise and detail the changes. One of the new features is the recognition of perspective projection stereochemistry.
<repository> <url>http://www.ebi.ac.uk/intact/maven/nexus/content/repositories/ebi-repo/</url> </repository> ... <dependency> <groupId>org.openscience.cdk</groupId> <artifactId>cdk-bundle</artifactId> <version>1.5.9</version> </dependency>
Monday, 1 December 2014
Currently there are a few limitations that we'll get out the way:
- Only generation of the CDK ECFP4 is supported and at a folded length of 1024, this should give a close approximation to what Matt used in MongoDB (RDKit Morgan FP). Other fingerprints and foldings could be used but the generation time of path based fingerprints in the CDK is currently (painfully) slow.
- Building the index is done in memory, since 1,000,000x1024 bit fingerprints is only 122 MiB you can easily build indexes with less than 10 million on modern hardware.
- During index searching the entire index is memory mapped, setting the chunks system property (see the GitHub readme) will avoid this at a slight performance cost.
- Results return the id in the index (indirection) and to get the original Id one would need to resolve it with another file (generated by mkidx).
- Index update operations are not supported without rebuilding it.
$ ./smi2fps /data/chembl_19.smi chembl_19.fps # ~5 mins $ ./mkidx chembl_19.fps chembl_19.idx # seconds
fpsscan does a linear search computing all Tanimoto's and outputting the lines that
are above a certain threshold. The
toper utils use the index, they either filter for similarity or the top k results. They can take multiple SMILES via the command line or from a file.
$ ./fpsscan /data/chembl_19.fps 'c1cc(c(cc1CCN)O)O' 0.7 # ~ 1 second $ ./simmer chembl_19.idx 0.7 'c1cc(c(cc1CCN)O)O' # < 1 second $ ./toper chembl_19.idx 50 'c1cc(c(cc1CCN)O)O' # < 1 second (top 50)
Using the same queries from the MongoDB search I get the following distribution of search times for different thresholds.
|Threshold||Median time (ms)|
It's interesting to see that the times seem to flatten out. By plotting how many fingerprints the search had to check we observe that below a certain threshold we are essentially checking the entire dataset.
The reason for this is potentially due to the sparse circular fingerprints. Examining the result file (see the github README) we can estimate that on average we're calculating 23,556,103 Tanimoto's a second. This also means that retrieving the top k queries isn't bad either. For example 10,000 gives a median time (Code 3) of 72 ms.
$ ./toper chembl_19.idx 10000 queries.smi
Next I'll look at some like-for-like comparisons.
Wednesday, 26 November 2014
The similarity searching partially uses the "Baldi" algorithm with some extra tweaks based on checking rare bits. The Baldi method is nicely summarised along with others by Tim Vandermeersch in his post on Fingerprint Searching Using Various Indexing Methods. As is noted by Tim, it can be improved upon.
Anyways, I had an implementation of a memory mapped Baldi index lying around, there is also one in the OrChem database cartridge. I prototyped the implementation back in April and was/is part of a "nfp" (new fingerprint) module for CDK. I've now put the code on a GitHub project (github/johnmay/efficient-bits/fp-idx) and will do some benchmarking to see how it does.
My feeling is that the very simple (it's about 100 lines) memory mapped index can give competitive performance on small datasets (<5 million entries).