Medical Bioinformatics and e-Bioscience

e-BioScience News

January 18, 2012

Hunting for viruses: finding a needle in a haystack. Published at EGI and iSGTW.

Illustration: wikicommons / CDC

November 25, Science Park Amsterdam

The e-bioscience group and the SHIWA project organized a workshop on "workflow interoperability" at the Science Park.

More info.

November 16, Den Haag

The ebioscience group demonstrated the e-bioinfra gateway at the ICT Delta 2011 conference.

This event also marked the official start of the COMMIT program, in which the AMC participates as partner of the Biobaking with imaging project.

June 5-8: In London

Early June was busy for the e-Science group because we participated in three events in London. The first was the pre-kickoff of the SCI-BUS project (SCIentific gateway Based User Support), where partners presented their ambitions regarding the development of scientific gateways for various user communities. The second was PUCOWO, the workshop of the P-Grade platform user community. The third was IWSG-Life, the workshop on Scientific Gateways for Life Sciences. Silvia presented talks in the first two, and Shayan presented a talk (via skype) in the third. His talk motivated an entry on the EGI blog.

June 1: Talk about Desktop Computing Grid at the AMC

Speaker: Daniel Westerbeek

Title: Setting up a desktop computing grid at the Academic Medical Center of Amsterdam

Date: wednesday 1 june

Room: CZ-2

Time: 13:30

Abstract: Biomedical research demands increasing amount of computing capacity for data analysis. At the Academic Medical Center of the University of Amsterdam (AMC) various experiments (e.g. DNA sequencing) already face capacity problems, requiring new solutions. However, many of them use private data that preferably should not leave the secured network of the hospital. At the same time, about 8,000 windows desktops are available at the AMC. Therefore, a local desktop grid seems to be an attractive option to leverage the existing computing capacity and prevent sending private data outside the AMC firewalls. The project tries to answer three research questions: whether it is possible to set up a local desktop grid at the AMC, what the value of the local desktop grid could be, and how acceptable such a desktop grid would be. With the support of the EDGeS project, a pilot desktop grid has been set up at the AMC, and this was used to evaluate the benefits. Moreover, a survey was performed to assess the acceptability. The obtained results indicate that it is possible to set up a local desktop grid at the AMC and that it could be beneficial, but only if certain conditions are met. Survey results indicate that doubts about a desktop grid are common, but that a desktop grid could be useful to them. The IT department is open to the idea, but also expresses doubts. These doubts have to be discussed with the IT department and taken away among other AMC employees before a local desktop grid would be acceptable.

April: Virus discovery on the Dutch Grid

Since March 2011 researchers of the Virus Discovery Unit, Laboratory of Experimental Virology of the Department of Medical Microbiology have been using the e-bioinfra gateway to perform the analyses of high throughput sequence data on the DutchGrid.

In total 1,021,470 reads were analysed multiple times via the BLAST tool, taking around 100 computing days.

Currently a group of 7 researches are involved in the analyses, being supported by Michel de Vries (Virus discovery unit) and Barbera van Schaik (KEBB - Bioinformatics laboratory).

April 11-14: AMC @ the EGI User Forum

The AMC actively participates in various sessions of the EGI User Forum in Vilnius:

• SHIWA project:
  • Multi-workflow systems interoperability: the SHIWA platform
  • Workflow interoperability solutions with case studies
  • hands-on tutorial about the SHIWA Simulation Platform
  • presentation by Vladimir Korkhov Coarse Grained Interoperability of Heterogeneous Grid Workflow Systems
• Poster by Barbera van Schaik: Large DNA re-sequencing experiments on the Dutch BiG Grid infrastructure
• Presentation by Mark Santcroos Worker Node Software Management: the VO perspective
• Presentation by Ammar Benabdelkader Getting a grip on the grid: A knowledge base to trace grid experiments
• Presentation by Vladimir : desktop grid EDGI workshop

March 16-18: Daniel Farkas visits the AMC

On wednesday March 16, thursday March 17 and friday March 18 Daniel Farkas from the University of Westminster, London, visited the eBioScience group to provide support at the pilot Desktop Grid project. He completed the deployment of the BOINC server and the software BWA, which performs sequence alignment on the AMC desktop grid test bed. See this page for more information.

March 10: last minute: talk about caBIG in the AMC

Dr. Alfredo Tirado-Ramos (Emory University School of Medicine, Atlanta, US) is visiting Amsterdam this week. He kindly agreed to present a last-minute talk at the AMC about his experience in setting up an infrastructure for translational research based on the Cancer Biomedical Informatics Grid (caBIG) platform

Date: 10 March
Time: 14-15h
Place: AMC
Title: Translational Informatics: a Couple of Case Studies in Cancer and Infectious Disease Research

Enhancing existing translational informatics infrastructure is becoming one of the top priorities to support overall biomedical research strategies for, e.g., biomarkers, biobanking and sample tracking, drug discovery, clinical trials, statistics, data management and analysis, translational science, and imaging and nanotechnology. In this talk we discuss a couple of use cases in translational informatics research targeting areas of relevance: large scale deployment of caBIG infrastructure, particularly caTissue and caArray, and joint work on HIV-related malignancy, particularly Lymphoma, between WCI and the Emory Center for AIDS Research, in the context of the Atlanta Clinical and Translational Science Institute.

Speaker:
Alfredo Tirado-Ramos, Ph.D.

Assistant Professor
Wallace H. Coulter Department of Biomedical Engineering
Georgia Institute of Technology and Emory University School of Medicine

Senior Scientist
[[http://cci.emory.edu/cms/index.html][Center for Comprehensive Informatics], Woodruff Health Sciences Center
Emory University

Feb 25: MR data analysis made easy

Talk Matthan Caan (Radiology) & Shayan Shahand (Bio-informatics) at the fMRI user's group meeting at the AMC

Date/Time: February 25th, 16.30-17.30
Location: AMC Psychiatry Department
Room: PA0-126

MR data analysis can be a complicated task. Each modality (e.g. T1, DTI, fMRI) requires a different approach. Numerous software packages are available that one has to become familiar with. Finally, data processing can be a lengthy procedure. We will present recent advances in grid technology, that facilitate 0;/script>data analysis. We will focus on Freesurfer segmentation and DTI registration. Shayan will present the e-bioinfragate, through which researchers are granted easy access to the grid.

Feb 2-3: Dr. Tamas Kiss visits the AMC

Dr. Tamas Kiss visited the AMC in Amsterdam on February 2-3 for collaboration with the e-Bioscience group of the Bioinformatics Laboratory of the KEBB on the area of Desktop Grid computing. This is an activity of the European Desktop Grid Initiative (EDGI) project.

See more details here.

Jan 18: Beta-release e-BioInfra Gateway (e-bioinfragate)

The new web interface to the e-BioScience Infrastructure (e-BioInfra) has been released! The e-BioInfra Gateway (e-bioinfra gateway) enables users without grid expertise to perform data analysis on the Dutch Grid from a web interface. Currently two neuroimage analysis programs are available to enhance and enable neuroscience research at the AMC. . The e-bioinfragate is developed and maintained by the e-BioScience group of the AMC with support from the AMC ICT innovation fund and the BiGGrid project.

Visit the e-bioinfra gateway and check the documentation for more information. Your feedback is very important to us!

Jan 10: “General Executor” helps researcher to run arbitrary software on the Grid

In an attempt to make the grid more accessible to biomedical researchers, we developed a General Executor (GE) for the grid. By using the GE the researcher can him/herself send an executable (including all needed libraries) with the input data to the grid, and specify the command needed to be performed on the data. GE takes care of running the command on the grid and storing the results where the researcher can find them later. The VBrowser is used as interface.

Dr. Matthan Caan, a researcher of the Radiology department of the AMC is currently using the GE to run registration of brain images (DTI) on the Dutch grid. More information about the GE can be obtained on here.

2010

Dec: New e-BioInfra server operational at the AMC

The new e-BioInfra server has been successfully deployed at the AMC premises. This server hosts all services of the e-BioInfra, for example web applications, data transfers and grid workflow execution and makes them available under the new ebioscience.amc.nl subdomain. This was a long term project involving the e-bioscience group of the Bioinformatics Laboratory and the AMC ICT department (ADICT). The new server is located at the De-Militarised Zone (DMZ) of the AMC network, which enables the communication with external systems like the Dutch e-Science Grid. The server is currently operating in testing phase and is expected to be put into full production in February.

Nov: Brain image segmentation in production

The department of radiology started to use the grid implementation of the image segmentation software Freesurfer. In a first attempt, a dataset containing 33 scans, each between 6 and 32 MB, were processed in parallel on an average of 20 hours for each scan (from 11 hours to 30 hours). The complete image analysis was completed in less than 36h. Before the grid implementation, the department of radiology processed the datasets sequentially, taking almost one day per scan and 33 days in total. Using the grid it took around one day. This work will help Anne Rienstra from the department of radiology in her PhD research on the changes over time in Alzheimer's disease and Mild Cognitive Impairment.

“Yet another grid success!” commented Matthan Caan from the department of radiology.

Oct: Systems Neuroscience Practical Course

In a 3-session practical course at the radiology department about "Systems Neuroscience" a group of 6 students used the DTI Atlas Web application. The students divided into 3 groups and each group calculated an Atlas out of 22 subjects using the DTI Atlas Web application concurrently. Each set of Atlas generation workflows took on avarage 3:30h of wall clock time, consumed 472 MB of files as input, generated 4512 MB of files as output, and used around 60 CPU hours.

The DTI Atlas Web application provided in this course was using the data staging mechanism to provide an easy-to-use fast data movement to/from the Grid. The complexity of the underlying Grid framework is completely shielded from the users. Indeed one of the students said: "I heard that the Grid computing is complicated and difficult, but in this course the easiest part was the Grid computing part: just a few clicks on the Web application and after some time when we saw "All done! no errors" on the report page we could get our results back and analyze them."

Oct: Matthan Caan defends PhD research

Matthan Caan (AMC Radiology and one of the valuable collaborators of the e-bioscience group) defended his PhD thesis on Oct 11 in Delft. His research was financed by the Virtual Laboratory for e-Science project (VL-e), which provided him the opportunity (and duty) to explore the added value of grid computing for in medical imaging science. Matthan's work was concentrated on the improvement of methods for the analysis of Magnetic Resonance Imaging (MRI) Diffusion Tensor Imaging (DTI), which only recently is being applied for clinical applications in neuroimaging. His work improved the details in which nerve bundles can be reconstructed from the images, contributing to development of markers of brain tissue degeneration, which attracted some media atten tion. The computations needed to develop, evaluate and optimize the new methods were performed on the Dutch grid using the e-Bioinfra platform. See more.

Sept: Brain image segmentation on the Dutch Grid

A new medical imaging application can now be executed on the grid: the freesurfer brain segmentation tool. This is the result of joint work of Yassene Mohammed with Matthan Caan and Paul Groot from the Radiology Dept. Now the freesurfer can be executed simultaneously for various scans, reducing the time needed to perform image analysis in neuroimaging experiments.

Sept: AMC at the BiGGrid User's Day

Mark, Barbera and Angela represented the AMC e-bioscience group at the BiG Grid user's day on September 14, 2010. Barbera presented a demo, and Mark presented a talk and participated in the closing panel. This generated a news item at the GridForum.NL newsletter.

Sept: Talk at the Science Park

The AMC e-BioScience group and BiGGrid hosted a talk at the Science Park in Amsterdam on 21 September 2010.

A Fresh Perspective on Distributed Scientific Applications and Cyberinfrastructure

Shantenu Jha

It is generally accepted that the ability to develop large-scale distributed applications that are extensible and independent of infrastructure details has lagged seriously behind other advances in cyberinfrastructure. Even accounting for the fact that developing applications for distributed infrastructure is a difficult undertaking, there are suspiciously few novel and extensible distributed applications that utilize multiple production distributed resources. The reasons are complex and defy over-simplication. As the sophistication and scale of distributed infrastructure increases, the complexity of successfully developing and executing such applications will increase in both quantitatively and in qualitatively newer ways. Against this backdrop, we will introduce the IDEAS design objectives -- Interoperabiltiy, Distributed Scale-out, Extensibilty, Adaptivity and Simplicity, for developing distributed applications. We then briefly outline SAGA -- which provides a simple, high-level POSIX-inspired API to the most commonly required distributed functionality independent of the underlying distributed environment details. We will discuss how SAGA facilitates the development of applications that can adhere to the IDEAS objectives. Finally, we will highlight several examples of how SAGA, and simple extensible abstractions built using SAGA, have been used by a range of different applications, tools and frameworks to effectively develop and execute distributed applications on a variety of distributed infrastructure, and thus provide a fresh perspective to developing distributed applications by addressing many limitations of the traditional modes of development.

Shantenu Jha is the Director for Cyberinfrastructure Development at the CCT, and a Research Professor in Computer Science at Louisiana State University (LSU). He is also a theme-leader at the e-Science Institute, Edinburgh and a Visiting Researcher at UC-London. His research interests lie at the triple point of Computer Science, Cyberinfrastructure Development and Computational Science. He is a co-PI for $2.6M NSF award that enables LSU/LONI participation as a TeraGrid Resource Provider. Shantenu leads the SAGA project and is currently working on co-writing a book on "Abstractions for Distributed Applications and Systems: A Computational Science Perspective" to be published by Wiley in 2010.

Aug: Script helps porting applications to the Dutch Grid

A Perl script has been developed by Angela Luyf to generate description files that facilitates porting an application (linux executable) to run as a (part of a) workflow with MOTEUR. This is a useful tool for anyone who is familiar with the syntax of the GASW configuration file. The command line script interactively obtains from the user basic information about the application (parameters, files, executable, etc) and automatically generated the XML files that describe the application as a service (GASW) and a workflow.

more information

microRNA data analysis can now be run on the AMC Life Science Grid cluster

Marcel Willemsen, Angela Luijf and Silvia Olabarriaga from the Bioinformatics Laboratory successfully ported a software application (CORONA) for the analysis of microRNA data to run on the AMC computing cluster that is part of the Dutch Life Science Grid (http://www.biggrid.nl). CORONA was developed by Applied BioSystems (currently Life Technology) and is used as part of a collaboration with Nick Schopman (Laboratory of Experimental Virology, Department of Medical Microbiology) for the analysis of microRNA data. Recently, cellular miRNAs were shown to target viral mRNAs transcripts. Here, we studied miRNA expression profiles in HIV-1 infected T-cells using the SOLiD ultra deep sequencing method. We show that HIV-1 production significantly affected the expression of several miRNAs and found that some of these miRNAs had potential target sites in the HIV-1 genome. It was initially installed on the computing cluster that is used to process raw data acquired by the Solid sequencer at the AMC genomics facility headed by Frank Baas. However, this cluster didn't provide the compute capacity that was required for performing the data analysis on a regular basis. With the installation of CORONA on the AMC cluster, Marcel Willemsen can now run the analysis regularly, without being constrained by other users and applications of the Solid computing cluster. The use of high performance and GRID computing is increasingly used for AMC research through the development of expertise, methodology and computational infrastructure (e-BioInfra) in the e-Bioscience group headed by Dr. Silvia Olabarriaga from the Bioinformatics Laboratory of the KEBB.

See more information here.

2009

Closing Workshop of the VL-e Medical project

The e-BioScience group organized the closing workshop for the Medical subprogram of the VL-e project. (more info)

Tristan Glatard at the Science Park

On 26 August 2009 Tristan Glatard presented a talk at an extra CT-Grid meeting (slides here)

My after-Dutch life: How much I miss grid.support ;-)

This talk will present our recent design, implementation, application porting and experimental efforts conducted at Creatis (Lyon, France) related to VL-e medical software. We will first present an integration of DIANE pilot jobs to improve quality of service in the biomed VO. Then, a service performing site monitoring and pre-selection will be described and experimental results will show how it complements pilot jobs. Besides, an interface of VL-e medical software with the ARC middleware will be presented, allowing application workflows to run both on EGEE/gLite and NorduGrid/ARC resources. Finally, three application use-cases (namely FIELD ultrasonic simulation, GATE radiotherapy simulation and CAVIAR cardiovascular analysis) will be presented, illustrating current achievements, remaining technical issues and on-going optimizations.

About the speaker:

Tristan Glatard worked in the Medical subprogram of VL-e (VLEMED) in 2007-2008. He is currently a CNRS researcher at the Creatis-LRMN lab, working on grid computing for medical imaging applications.

Contact with speaker:

Tristan Glatard
CNRS researcher
University of Lyon ; CREATIS-LRMN
http://www.creatis.insa-lyon.fr/~glatard
+33 (0) 472 43 72 99 begin_of_the_skype_highlighting +33 (0) 472 43 72 99 end_of_the_skype_highlighting begin_of_the_skype_highlighting +33 (0) 472 43 72 99 begin_of_the_skype_highlighting +33 (0) 472 43 72 99 end_of_the_skype_highlighting end_of_the_skype_highlighting

This event is jointly organized by VL-e, Big Grid and the e-Bioscience group of the Bioinformatics Laboratory of the AMC.

1 July 2009

The paper by Silvia Olabarriaga, Tristan Glatard, Andreas Hoheisel, Aart Nederveen and Dagmar Krefting entitled Crossing HealthGrid Borders: Early Results in Medical Imaging was presented at the Healthgrid Workshop 2009 in Berlin. The paper obtained the third place in the "Best paper award" competition, which was based on votes of a jury of experts during the conference. In 2008 this competition was won by the same group from the e-BioScience group of the AMC with another paper (see 2008 paper abstract).

Abstract

Image analysis has been strongly present from the beginning in several healthgrid initiatives, and today we find many imaging projects with successful grid implementations and developments. An example is the analysis of functional MRI data on grids, which has been successfully done by several projects and that could be of interest for others. However, crossing the borders of existing grids is not trivial because the infrastructures being created for these projects differ, each adopting a (slightly) different software stack. This paper describes our early attempts to cross the borders between the German and Dutch grid infrastructures for medical imaging, motivated by a true wish to share expertise about fMRI analysis on grids between these two communities. We describe how we used off-the-shelf, production-level, grid technology to implement supporting mechanisms for cooperation in fMRI at several levels (users, data, software, workflows and computing resources). This simple exercise provided us valuable insights into the problems of crossing the borders of real grids from a user's perspective. Besides technical aspects, we observed that security and usability are very important for the success of inter-operation of Healthgrid.

This paper presents the results of a collaboration between the AMC, the Charite Hospital in Berlin, CREATIS in Lyon (FR) and the Fraunhoffer Institute in Berlin.