Chemical decision support with Chemspider and ChEMBL-RDF

Bioclipse Decision Support is a system where users can draw or import chemical structures, execute multiple predictions or database lookups, and visually interpret results. This was previously demonstrated for chemical liability assessment, and for running OpenTox predictions.

In the recent days I have implemented a module for adding a ChemSpider SimilaritySearch complemented by ChEMBL lookup via the ChEMBL-RDF data maintained by Egon Willighagen. In short this is how it operates:

1. Query ChemSpider using the SimilaritySearch via the Web API (SOAP).
2. For a maximum of 15 nearest neighbors in Chemspider, query ChEMBL-RDF for interactions, and present information about assays, targets, and interaction values.

Below you can find a screenshot of how it looks like in action. Now, the ChemSpider similarity search is rather slow (>30 s) but then again it takes a while to query 27 M chemical structures.

Note: This work was done against a development version of Chemspider Web API. You also need a security token in order to be able to query Chemspider; it's free and can be obtained here. When you have a security token, paste it into the Chemspider section of the Bioclipse preferences. The Chemspider + ChEMBL model for Bioclipse Decision Support will be part of the 2.6 release of Bioclipse, which we hope to have in place within a few weeks.

Thanks to Egon Willighagen (ChEMBL-RDF) and Valery Tkachenko (Chemspider) for aiding me in this work.

Postdoc position in predictive toxicology at Uppsala University

Postdoc position in cheminformatics, bioinformatics, or computer science - applications in predictive toxicology

We have an open postdoc position in the group of Pharmaceutical Bioinformatics at the Department of Pharmaceutical Biosciences, Uppsala University, Sweden.

The successful applicant will conduct research on data interoperability for predictive toxicology, and especially design and implement an infrastructure consisting of a database and user interfaces for data and predictive models in toxicology. Of particular interest will be to merge chemical and biological data within a semantic framework, and link toxicity data to genomics and metabolomics data (toxicogenomics) with a connection to the Bioclipse framework (www.bioclipse.net). PhD degree or equivalent scholarly competence in a relevant branch of chem/bioinformatics or computer science and a strong interest in informatics and data integration is required. Required competences include web programming, databases, and working knowledge in Java. Experience with linked data is desirable.

Deadline for application: May 9th , 2012.

Link to job ad and application form: http://www.uu.se/jobb/others/annonsvisning?tarContentId=186211&languageId=1

Bioclipse meets OpenTox

Barry Hardy wrote a very nice post about when he encountered the Tamboti Tree, a very toxic tree, and his experience when using Bioclipse and OpenTox to assess its toxicity. Read the full article and watch the movie here: http://barryhardy.blogs.com/theferryman/2011/03/the-tamboti-tree-use-case-bioclipse-meets-opentox.html

Presenting at Society of Toxicology 2011

Tomorrow I will present a poster at the Society of Toxicology Annual Meeting 2011 in Washington DC. With over 7,500 participants, this conference is rather large. My poster has the title: "A flexible method for building and using predictive models applied to safety endpoints" and mainly described the Decision Support feature for Bioclipse. Anyway, this is what my poster looks like:

Web seminar on Bioclipse

Ola Spjuth of the Bioclipse project will give a web seminar on Mar 29th 16:00 GMT entitled: In-Silico Prediction of Drug Safety and Metabolites Using Bioclipse. The seminar is free of charge but requires registration.

Bioclipse 2.4.1 now on update site

A new version can be found on the update site. The new versions of the features contain the latest fixes. New in this version is support for Jasper reports which are so far used by Bioclipse Decision Support and Bioclipse Brunn for generating printable reports.

To update go to Help -> Software Updates... and select the features you want to install.

Known problem: The splash screen erroneously shows version number to be 2.4.0 after update. (Technical information can be found at bug: 2084 in our Bugzilla)

Bioclipse 2.4 released

The Bioclipse team is proud to announce the release of Bioclipse version 2.4. The release contains various new features and bug fixes in cheminformatics and drug discovery, including improved QSAR functionality, site-of-metabolism prediction, semantic web functionality, browsing of large compound collections, editing of chemical structures, and numerous bug fixes.

Bioclipse 2.4 is available for 32 and 64 bit versions of Mac OS X, Linux, and 32 bit version of Windows (Bioclipse for 64 bit Windows is currently unavailable, but will be provided as soon as a native Standard InChI is available for 64 bit Windows).

• Download Bioclipse
• Getting started guide
• Planet Bioclipse integrates blogs related to Bioclipse.

Bioclipse 2.4.0.RC3 is here

A new release candidate is out and can be found at the usual site for release candidates:

http://pele.farmbio.uu.se/bioclipse-devel/

I have a good feeling about this one. I think Bioclipse 2.4 is really close now.

And as usual, if you download and try the release candidate of course you already know that we love to get bug reports in our Bugzilla. :)

Bioclipse 2.4.0.RC1 is out

A new release candidate is out and can be found at the usual site for release candidates:

http://pele.farmbio.uu.se/bioclipse-devel/

What is new?
Among the main news are:

• New molecules table.
• We are now using Java 1.6
• Eclipse 3.5.2
• CDK 1.3.5

And of course bug fixes and probably a lot of other stuff which I don't know about. If you download and try the release candidate of course you already know that we love to get bug reports in our Bugzilla.

Bioclipse is finalist for the Eclipse Community Awards 2010

Bioclipse is one of three Finalists for the Eclipse Community Awards 2010 in the category Best RCP Application. I am looking forward to going to EclipseCon, it will be my first visit. I have also submitted a poster, and hope it will be among the selected ones (see below or the submission abstract). UPDATED: My poster is now accepted!

Below is the screencast which I constructed for the Community Awards submission:

Bioclipse 2.2 released

The Bioclipse Team is proud to announce the release of Bioclipse 2.2.0. The new version includes, apart from numerous bug fixes, the following highlights:

* Cheminformatics, with a pure SWT-based chemical 2D editor (JChemPaint) and a lazy-loading molecules table.
* QSAR, supporting local, REST, and XMPP services
* MetaPrint2D for interactive site-of-metabolism prediction for chemical structures
* StructureDB and VScreen: A chemical database with virtual screening functionality
* The new Decision Support feature with graphical reports using BIRT
* Semantic web features
* Bioinformatics, with the new Sequence Editor and sequence alignments via the Kalign Web service (Experimental)

A screenshot from Bioclipse with the MetaPrint2D feature showing predicted sites of metabolsim for a set of drugs in the MoleculesTable.

Note that Bioclipse 2.2.0 requires a fresh download, i.e. it can not be upgraded to by using the software update functionality. A small installation guide is also provided, but the main documentation for Bioclipse is available from help.bioclipse.net; the same information is also available from within Bioclipse from the menu Help > Help Contents. For general questions there is the bioclipse-users and bioclipse.devel mailing lists.

Links:

• Download Bioclipse 2.2.0.
• Bioclipse help (online version)
• Bioclipse website
• Bioclipse 2.0 announcement
• Installation guide
• Report a bug
• Mailing lists

Congratulations doctor Ola Spjuth

Ola Spjuth, the mastermind behind the realization of Bioclipse is now a doctor. These last days he and his best friend Martin Eklund went through the process. Olas thesis can be found here and more importantly I have it all documented in pictures on flickr.

Download PDBs with Bioclipse

It is easy to download bioinformatics resources into Bioclipse using the WSDbfetch Web service at the European Bioinformatics Institute (EBI). From console you can use the method:

ws.downloadDbEntry(String db, String query, String format)

From the GUI you can use menu: File > New... and select the Wizard Download > Query WSDbfetch at EBI. Click next and see the many available databases in the dropdown list Supported databases:

Try for example to download the sequence "NM_210721" from the "refseq" database.

If you'd like to download proteins in PDB format, there is a convenience wizard available in the New Wizards dialog available from the menu File > New...


Clicking Next allows for inputting a comma-separated list of PDB IDs, try for example "1ale,2pdz". Clicking next downloads the file to the selected folder in the Navigator, or to the Virtual project if nothing is selected. Below is visualized 2PDZ in the Jmol Editor.

Align sequences with the Kalign Web service

A recent feature addition to Bioclipse is the ability to align protein sequences using the Kalign Web service available from the European Bioinformatics Institute (EBI). Simply select two protein sequences in the Navigator (currently the FASTA format is only supported but more formats are in the pipe), right-click, and select Align > Align using Kalign.


This compiles the sequences into the input format that Kalign expects, and sends it via SOAP to the Kalign Web service at EBI. The results is stored as a file in the same folder as the aligned resources, and opened in the SequenceEditor (note the new Wrap feature available from the toolbar).


The feature is available from the latest development version of Bioclipse.

Bioclipse 2.1.1 devel release

A new development release is now available for download at http://pele.farmbio.uu.se/bioclipse-devel/. Highlights of the development update site (http://update2-devel.bioclipse.net) are:

• Bioinformatics, with the new Sequence Editor now capable of wrapping sequences
• Bioclipse Reporting (BIRT)
• QSAR, supporting local, REST, and XMPP services
• Semantic web features (RDF and Pellet)
• StructureDB: A chemical database for Bioclipse
• The new VScreen feature for Virtual Screening
• The new Decision Support feature
• The new MetaPrint2D feature

Note that the development versions are not compatible with the stable 2.0.x release branch which is available from Sourceforge. This update replaces the 2.1.0 devel release, and the devel update site is only compatible with the 2.1.1 release.

Figure: Screenshot teaser from the Bioclipse 2.1.1 devel release showing the Decision Support feature in action for a collection of molecules.

Virtual Screening for Bioclipse

I recently created a Bioclipse feature to filter chemical structures in a StructureDB database (the new database for chemical structures in Bioclipse). The new plugin is called VScreen, and aims at being a future Virtual Screening feature. Currently the feature is in very early proof-of-concept state with only two simple filters, Molecular Weight and XlogP. Plans include to create an extension point for filters that plugin can extend to contribute new filters.

Currently, the only way of interacting with the plugin is via Bioclipse scripts. A sample usage is available below in a Gist:



A screenshot from Bioclipse is shown below:

Bioclipse 2.1.0 development release

Today the first Bioclipse development version (versioned 2.1.0.v20090807) was released. A development update site (http://update2-devel.bioclipse.net) is packed with features to test (for example the new Bioinformatics feature, early versions of the StructureDB, QSAR, and RDF features). Note that the development versions are not compatible with the stable 2.0.x release branch which is available from Sourceforge.

There will be frequent updates to the update site for this, towards the next stable version 2.2.0. For more information and downloads of this development release, see http://pele.farmbio.uu.se/bioclipse-devel/.

Bioclipse 2.0 released

On behalf of all Bioclipse developers I am happy to announce the release of Bioclipse 2.0. Bioclipse is a free, open source workbench for the life sciences that provides advanced functionality mainly in cheminformatics (bioinformatics is planned for version 2.1 later this summer). Some major components include a brand new chemical editor for SWT (JChemPaint), interactive 3D visualization of molecules (Jmol), a Molecules Table capable of reading large files, and a powerful backbone in cheminformatics provided by the Chemistry Development Kit (CDK) library.

Figure 1: Screenshot of Bioclipse showing editing of a chemical structure using the new JChemPaint editor.

Bioclipse is a Rich Client for the life sciences that provides the means to run and integrate algorithms and tools in disconnected state, while still taking advantage of remote services if a network connection is available. Built on the famous Eclipse framework, Bioclipse delivers a state-of-the-art plugin architecture which makes it possible to extend it in any direction.

Figure 2: Screenshot of the interactive 3D visualization of a protein using the integrated component Jmol.

All functionality in Bioclipse 2 is available from the GUI as well as a new scripting language based on Javascript. This allows for complete control of the workbench and functionality from scripts, which can be used to automate tasks or reproduce and validate scientific analyses.

Figure 3: All functionality in Bioclipse is available from an integrated scripting language based on Javascript.

Bioclipse 2 can be downloaded from Sourceforge, releases are available for all major platforms. There is an update site where users can install additional functionality (such as Speclipse) and data collections; this is available from the Bioclipse workbench under menu Help > Software updates.

A small installation guide is also provided, but the main documentation for Bioclipse is available from help.bioclipse.net; the same information is also available from within Bioclipse from the menu Help > Help Contents. For general questions there is the bioclipse-users mailing list.

All software contains bugs, and Bioclipse is no exception. However, in contrast to many commercial and closed source initiatives, open source projects generally have a faster bug fixing rate as well as more frequent releases. If you find bugs in Bioclipse, please report them on bugs.bioclipse.net. There is a list of intractable bugs on the Bioclipse development wiki, and also a convenience list for tracking known major bugs.

Bioclipse is an open development that welcome new developers with varying backgrounds. Developers hang out on daily basis on IRC (irc.freenode.net, channel #bioclipse), and can also be reached via the mailing list bioclipse-devel.

Thanks to all contributors who made this release possible!

Working with large SDFiles in Bioclipse

I decided to test the performance of Bioclipse 2 (current release 2.0.0RC5) for working with large structural files (SDFiles). I first loaded in the complete ChEBI (Chemical Entities of Biological Interest) which consisted of 13.486 chemical structures and has a file size of 54 MB on disk. This was very fast, Bioclipse indexed and opened the file in less than 4 seconds, and then continued to parse the properties in the background for another 4 seconds (but during this time it is possible to browse and work with the structures). The MolTable editor was very responsive and scrolls nicely.

Figure 1: Screenshot from Bioclipse with the entire Chebi SDF open.

To really push Bioclipse, a test file of the first 225.000 compounds in Pubchem were concatenated, resulting in an SDFile of size 1.1 Gb. Bioclipse creates an index of the file and opens it in 66 seconds. It then continues parsing the properties in the background, which takes another 78 seconds. The MolTable editor was still very responsive and scrolls nicely. Not bad for such a large file!

Calculating InChI on the >1Gb file on the open file in MoleculesTable (resulting in all InChI properties kept in memory) took 13.20 min. Trying to save the resulting file took 2 min 49 seconds for the first 20Mb, extrapolated to 2h and 20 minutes for the total (this forces a complete save of all chemical structures and a lot of swapping in and out from disc). Calculating the same InChi and saving to file but not opening it in MolTable first (avoiding keeping all properties in memory) took 20 minutes. What do we learn from this? Browse large files is fine, but if you want to manipulate them then, do this on the file directly without visual inspection.

As a side note: Handling large SDFiles is generally not a recommended solution. When StructureDB (a relational database for chemistry) is released for Bioclipse, we will see a dramatic performance boost when dealing with large collections of molecules.

Bioclipse 2.0 Release Candidate 5

Today, Bioclipse 2.0 Release Candidate 5 (versioned 2.0.0.RC5) was released with primarily a fix in the atom typing done when editing chemical structures, and a less stricter handling of SDFiles. The Bioclipse help is also available as standalone. The release requires a fresh download from Sourceforge, and we kindly ask beta-testers for bug reports on the bugs.bioclipse.net.