Real Estate

ABA-Cloud: support for collaborative breath research

Description
ABA-Cloud: support for collaborative breath research
Categories
Published
of 14
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  This content has been downloaded from IOPscience. Please scroll down to see the full text.Download details:IP Address: 60.12.218.137This content was downloaded on 18/10/2013 at 17:07Please note that terms and conditions apply. ABA-Cloud: support for collaborative breath research View the table of contents for this issue, or go to the  journal homepage for more 2013 J. Breath Res. 7 026007(http://iopscience.iop.org/1752-7163/7/2/026007)HomeSearchCollectionsJournalsAboutContact usMy IOPscience  OPEN ACCESS IOP P UBLISHING  J OURNAL OF  B REATH  R ESEARCH J. Breath Res.  7  (2013) 026007 (13pp)  doi:10.1088/1752-7155/7/2/026007 ABA-Cloud: support for collaborativebreath research Ibrahim Elsayed 1 , Thomas Ludescher 2 , Julian King 3 , Clemens Ager 3 ,Michael Trosin 1 , Uygar Senocak 1 , Peter Brezany 1 , Thomas Feilhauer 2 and Anton Amann 3,4 1 Research Group Scientific Computing, University of Vienna, W¨ahringer Strasse 29, A-1090 Vienna,Austria 2 Department of Computer Science, University of Applied Sciences, Hochschulstrasse 1,A-6850 Dornbirn, Austria 3 Breath Research Institute of the Austrian Academy of Sciences, Rathausplatz 4, A-6080 Dornbirn,Austria 4 University-Clinic for Anaesthesia, Innsbruck Medical University, Anichstrasse 35, A-6020 Innsbruck,AustriaE-mail: thomas.ludescher@fhv.at Received 25 September 2012Accepted for publication 27 March 2013Published 26 April 2013Online at stacks.iop.org/JBR/7/026007 Abstract This paper introduces the advanced breath analysis (ABA) platform, an innovative scientificresearch platform for the entire breath research domain. Within the ABA project, we areinvestigating novel data management concepts and semantic web technologies to documentbreath analysis studies for the long run as well as to enable their full automatic reproducibility.We propose several concept taxonomies (a hierarchical order of terms from a glossary of terms), which can be seen as a first step toward the definition of conceptualized termscommonly used by the international community of breath researchers. They build the basis forthe development of an ontology (a concept from computer science used for communicationbetween machines and/or humans and representation and reuse of knowledge) dedicated tobreath research.(Some figures may appear in colour only in the online journal) 1. Introduction In the field of breath research, the number of scientificstudies has increased considerably over the last few years dueto availability of novel scientific instrumentation providingimproved analytical methods for the investigation of volatilecompounds in exhaled breath. In addition, data processing andstatistical computations in breath research studies are creatingvast data stores that require new methods to organize theentire data life cycle of such studies. Typical measurementtechniques such as two-dimensional gas chromatography withthetime-of-flightmass-spectrometricdetection(GCxGC-ToF- Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and thetitle of the work, journal citation and DOI. MS) [1] or the proton-transfer-reaction time-of-flight massspectrometry (PTR-ToF-MS) [2] deliver huge amounts of raw data which have to be processed by software which will beoutdated in a few years time. Also, medical parameters suchas the ECG, ventilatory flow, or EEG streams during real-time measurements of exhaled breath in ergometer stresschallenges or in the sleep laboratory require considerablestorage capacity [3, 4]. Therefore, the documentation of the measurement process, the raw data, the processed data, themethod of data integration, etc, is an increasingly complextask. Hence, a simplified access to a complete breath researchstudy as well as its documentation and repeatability (even15 years after the completion) is a crucial factor to fulfil legalrequirements.Recently, ensembles of distributed resources or so-called clouds  have emerged as popular platforms for deploying 1752-7155/13/026007 + 13$33.00  1  © 2013 IOP Publishing Ltd Printed in the UK & the USA  J. Breath Res.  7  (2013) 026007 I Elsayed  et al data-intensive scientific applications. Clouds are nowconsideredastheparadigmforthenextgenerationofscientificcomputing and data management with the main advantage ineliminating the need for hosting expensive hardware.Within the ABA project (http://aba.cloudminer.org), we are developing a novel cloud-based information infrastructurefortheinternationalbreathresearchcommunity.TheResearchGroup for Scientific Computing at the University of Viennaand the University of Applied Sciences Vorarlberg haveimplemented and evaluated a first prototype of a cloud-basedinformationinfrastructurecalledABA-Cloud.Itdeliversservices for executing code in distributed environments andpreserving the algorithms for statistical data evaluation withinexisting problem solving environments (PSEs) such as Matlab[5], R [6] and Octave [7]. It also provides a framework, based on semantic web technologies to annotate breath researchstudies according to a predefined set of attributes. The firstprototype has been developed in collaboration with leadingbreath researchers from the Breath Research Institute of theAustrian Academy of Science. Currently, the prototype isbeing tested by a small core of breath researchers acting asearly adopters.One of the most important challenges we are addressingin the ABA project is to enable the full automatic repeatabilityof entire breath research studies. For this purpose, one needsto consider both data management (efficient storage of allrelevant datasets) and code execution (for data analysis andstatistical algorithms). ABA-Cloud integrates and furtherdevelops state-of-the-art approaches to address the entire datalifecycleintheprocessofperformingscientificstudies,aimingatfurtherenhancingreuseanddisseminationofbreathstudies.There are many web-based platforms available thataddress the dissemination of domain-specific research data.A pioneering example is represented by  PhysioNet  , whichis an online forum for the dissemination and exchange of recorded biomedical signals and open-source software foranalyzing them. It provides facilities for the cooperativeanalysisofdataandtheevaluationofproposednewalgorithms[8]. Another related approach is represented by  VectorBase ,which is a bioinformatics portal that focuses on storinggenomic and related data on invertebrate vectors that transmithumandiseases[9].Itoffersmultipleintegratedbioinformaticstools for analyzing and browsing genomic and relateddata. What makes ABA-Cloud unique, compared to above-mentioned examples, is its integration into the PSE usedby the researchers jointly working at a study and preparingappropriate presentations and publications. Moreover, ABA-Cloud provides not only tools for the dissemination of breathstudies but also for their repeatability during the 15-year legalperiod after the performance of a clinical study. By this, ABA-Cloud supports the application of statistical algorithms anddata analysis methods on a broader scale, e.g. benchmark datasets could be disseminated for a consistent evaluation of certain classification methods.The rest of this paper is organized as follows. In section 2,we discuss some motivating real-world usage scenarios inorder to present how the user will be using ABA-Cloud andprofitingfromitsservices.Insection3,wepresentanoverview of the ABA-Cloud system architecture including its servicesand discuss some related works. In section 4, we presenta first proposal of concept taxonomies for breath researchand highlight how different users will be interacting with thesemanticrepository,whichisakeycomponentinABA-Cloud.In section 5, two breath research studies conducted by early adopters from the Breath Research Institute of the AustrianAcademyofScienceusingABA-Cloudservicesarepresented.Finally, in section 6, we conclude this paper and in section 7, we briefly present our next steps. 2. Motivation and system usage ABA-Cloud provides interfaces to (a) publish entire breathresearch studies with all relevant data and to (b) submitkeyword-basedqueriesinordertosearchforstudiesconductedat a participating research center. Publishing a study canbe done from within a PSE such as Matlab [5], R [6] and Octave [7]. SearchingtheunderlyingspaceofdatathatcontainsallthepublishedbreathresearchstudiescanbedoneeitheragainfromwithinaPSEorviaspecificallydevelopedinterfacesintegratedinto the ABA-portal. The ABA-portal provides special toolsallowing us to search and browse for studies in ABA-Cloud.Figure 1 depicts a screenshot of the  ABA-Study Visualizer  ,which is a tool developed to visualize the most importantsemantic information about a single breath research study.Figure 2 shows a screenshot of the  ABA-Study Browser  , whichaims at visually browsing studies available in ABA-Cloud.SearchingtheABA-Cloudisnotlikesearchingadatabase,which is typically done with one-shot queries. Rather, it isan iterative process where breath researchers first submit akeyword-basedquery(e.g.monitoringacetoneandisoprene—PTR), then retrieve a ranked list of studies matching thekeyword query and—based on further selections made bythe user—may explore selected studies in more detail withall related datasets and semantic information connected to thestudy.Inthefollowing,weoutlineseveralrealusagescenarios,eachwithabriefmotivationonhowscientistscanbenefitfromthe information infrastructure provided by ABA-Cloud.  Automatic reproducibility of breath research studies.  In adynamic research environment with scientists continuouslyentering and leaving research groups, it will hardly bepossible to retrieve all relevant data of a specific studyonce the responsible scientist who conducted the study hasleft the group. In fact, all information about the study thatis left back at the research group is stored either withinscientific publications, technical reports or other kinds of documentations written by the corresponding researcher.The information represented in such documents howeverdoes not allow us to reproduce the study. Conversely, if the scientists have published the study using the servicesprovidedbyABA-Cloud(e.g.usingtheMatlabtoolbox,whichprovides a template structuring and documenting a breathstudy according to predefined attributes), it is accessible inthe underlying space of data together with all relevant datasetssuch as the input dataset, the analytical methods applied 2  J. Breath Res.  7  (2013) 026007 I Elsayed  et al Figure 1.  Screenshot of the ABA-Study Visualizer. and scientific publications related to the study. In addition,semantic information is available making the study bettersearchable and retrievable.  Reusing certain classifiers for different studies.  Consider alarge number of breath research studies from various researchcenters participating in the ABA-Cloud. Search servicesprovidedbytoolsviatheABA-portalallownotonlytodiscoverentirestudiesbutalsoenabletheirreuse.Onceastudyhasbeendiscovered by a breath researcher using e.g. the ABA-StudyBrowser, it can easily be downloaded to the local computer(given that the requesting scientist has the needed accessrights) either from the portal itself or directly from within thePSE. Thus, certain classifiers could be reused from differentstudies. 3  J. Breath Res.  7  (2013) 026007 I Elsayed  et al Figure 2.  Screenshot of the ABA-Study Browser. Preparation of material for quick and easy creation of manuscripts.  Anotherimportantusagescenarioisasfollows:consider a research group, where several scientists arecollecting breath samples and apply analytical methodsincluding a statistical analysis. Once these data are madeavailablewithinABA-Cloudusingitspublishingservices,e.g.using the Matlab toolbox, they can be searched, analyzed andvisualized using tools provided by the ABA-portal; thus, allrelevantmaterialforthequickandeasycreationofmanuscriptscan easily be prepared. 3. System overview ABA-Cloud is presented as a system that includesmultiple geographically distributed platforms, so-calledABA-platforms. Each ABA-platform represents a breathresearch group or lab, typically employing several personsincluding senior and post-doc researchers, master students,administrators and technicians. These persons are acting aslocal users on their own platform, while they are guest users inother platforms. Access rights can be defined by the research 4
Search
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks