Supplementary Material

Transcription

1 Supplementary Material Victoria Martín-Requena, Javier Ríos, Maximiliano García, Sergio Ramírez and Oswaldo Trelles Bioinformatics and Information Technologies Laboratory ( University of Malaga, Spain Index 1. jorca Quick start: installation and a simple exercise Quick Tour on jorca features Guided exercises on web service pipelining Using Web Services from the same protocol and repository Using Web Services - different protocols and repositories Using Web Services in a GRID environment Executing R-BioMoby Web Services Guided exercises on data standardization Comparison of Web Services clients Integration using the Modular API Model View Controller Pattern Ongoing work Feedback Supplementary Material 1

2 1. jorca jorca is a desktop client able to efficiently integrate different type of web-services repositories mapping metadata over a general definition to support scalable service discovery and to achieve flexible inter-communication between tools. jorca manages repositories heterogeneity supported by the Modular-API that provides a uniform view of metadata (e.g. GRID-based, WSDL-SOAP services and BioMoby), making the integration of bioinformatics Web-Services easier. jorca allows the execution of different web-services invocation protocols and includes features to cover a broad range of user skills: drag-and-drop edition styles, easy incorporation of viewers for data types, custom favourite s section that allows the user to freely define a customized environment, on-line updated documentation, etc and it is fully configurable by means of default parameters in text-files (e.g. initial-links, look & feel, etc). Several methods are provided to help user in discovering tasks to identify the appropriated data types or tools to solve specific problems: text boxes for fast keyword based searches, compatible webservices related to a type of data, advanced keyword searching including did you mean methods to manage misspelling words and endowed with user profile learning capabilities. An automatic interface builder provides a uniform look for service parameters and invocation tools, making easier the pipelining of tools by I/O compatibility. Advanced features such as mirroring or manual scheduling, asynchronous calls and execution tracing (editable log record of events) are also provided. Supplementary Material 2

3 2. Quick start: installation and a simple exercise Installation files are available at Java Virtual Machine version 6 is required. Java Installation instruction: Download the latest version for your OS. Create a new folder ($FOLDER) and uncompress the downloaded file. Double-click in jorca.bat (Win), jorca.sh (linux) file inside $FOLDER. In the first execution, jorca will request a HOME folder for the results files, choose the Guided-exercises folder inside $FOLDER. Supplementary Material 3

4 Once jorca s main window is opened, double click on the NCBIBlastp service in the favourite tree, the blast parameter information window will open. Double click Click on search for compatible Click on the search compatible object icon, wait for the results and click on myaaseqobject.xml. Now launch the tool using the Launch Service button. When the tab becomes green, you will have executed your first Web Service using jorca! Supplementary Material 4

5 3. Quick Tour on jorca features jorca is able to execute Web Services from different providers and protocols. jorca is a Desktop application that uses the local file system in the user machine to manage data files. The default folder can be easily changed in the configuration dialog. In the first execution, jorca program will display the tools available for the default repository (service provider). This default repository is also configurable. A predefined favourite tree has been included in this version to help getting started with Web Services. These tools can be found in favourite s area. Look and feel is also configurable: colours, icons, if trees are shown or not and how to display them. The user display is organized in three main areas. On the left-hand side the Control area that acts as a tools browser to locate the appropriate web service and invoke it (double click style). Data types and namespaces as well as a Favourite section is available and organized as browse-able trees, using the information about the repository ontology (Functional categories), ordered by authority or alphabetically. Supplementary Material 5

6 ToolBar and Menu Area Change repository: jorca can work with different Web Services repositories; this option allows selecting a new metadata repository. Reload: repository data is cached, use this option to refresh this data. Advanced search: use Magallanes ( search engine. Advanced configuration: configuration wizard to change default values. Execution log: view the executed task and their results, re-launch executed tasks and create SCUFL workflows. Guided tour: a link to guided exercises Workflow creation wizard: Magallanes workflow search tool Help: a link to jorca full help. Control Area This area displays the active repository tools (Web Services), data types (data) and namespaces (when available) organized as browseable trees, using the information about the repository ontology (Functional categories), ordered by authority or alphabetically. The favourites section is also available in this area. The display can be configured using the menu: hide trees, change configuration etc. The user can select an object or service by searching it in the tree or using the fast-search tool (text box in the upper position). When the user writes a word in the text box the matching services/data types will be highlighted automatically. A short description of each element in the tree is displayed in a tip when the mouse is over the name. To use a web service o create a data type double click or drag and drop it into the main Area. Supplementary Material 6

7 Display configuration The resources display mode can be changed from the Configuration menu. To switch among display modes use the display configuration. Favourites jorca has a Favourites feature to put tools or data types often used within easy reach. That way, the user doesn't have to remember or type anything. Just clicking the mouse twice and there you are. These tools are organized through their own categories/ icons/ name/ description. A predefined set of tools is supplied with this installation (the user can modify this content). tree. Using this button on the tool or data type panel it will be added to the favourite Supplementary Material 7

8 Control Area Services or Tools aim to process a given object (input data) in order to produce a new object (output result). The service behavior is personalized using parameters. This area is used for completing tools parameters, using the data-type editor, advanced search engines and in general, it is the main area of the application. Welcome page This tab provides some useful links in bioinformatics. These links can be easily configured using a configuration file. See full help at or click in the help button to get more information. Supplementary Material 8

9 Advanced Search Advanced Search icon opens Magallanes search GUI. Magallanes is full integrated in jorca giving the user the possibility to: Create a new data type, execute a tool (Web service) or compose pipeline of tools. More information is available at ( Supplementary Material 9

10 Service parameter composition The parameter composition window for a tool is displayed as a new tab in the Main Area. When the user clicks on a tool in the Control Area, a search result or drag a tool from the Control Area to the Main Area this tab will open. The advanced execution options (mirroring/asynchronism) are also available but they are initially hidden. Data-type parameter composition The parameter window for a data type is displayed as a new tab in the Main Area. The user must complete the parameter information to create a new object (a file with the data standardized) from a data-type (a definition about data) Supplementary Material 10

11 File System Area jorca uses this area to display the user files (starting in the HOME directory), an heuristic approach is used to identify the data-type that matches the file content. Some traditional file management abilities are available (creating new folder, refreshing data, navigation, and file deletion.) From this area, files can be used: a) as inputs for a compatible web service. b) to display the content using a predefined viewer. c) To be dragged into a web service parameter interface. In the picture, the AminoAcidSequence.xml file is going to be used as input for a compatible tool. Supplementary Material 11

12 4. Guided exercises on web service pipelining 4 guided exercises on Web Service pipelining will guide you through the use of jorca: Executing tools from the same protocol and repository Executing tools from different protocols and repositories Executing tools from in a GRID environment. Executing R BioMoby Web Services. The example files can be found at the same folder at jorca s home folder at start-up. (See Download and install jorca ). This package is composed by: o CTLA4_Human.txt Fasta sequence file o Q9M4M3_WHEAT.fas FASTA format file o GE-demo.dat Gene expression Matrix o R-data.txt Tab separated value data. o myaaseq.xml Aminoacid Sequence MOBY object. o Raw folder with the files for the guided exercises on data conversion. 3Seqs.fas 3 sequences in FASTA format PDB-raw.txt a PDB in plain text 1Seqs.txt 1 sequence Supplementary Material 12

13 Using Web Services from the same protocol and repository In this exercise: We will create an Aminoacid Sequence Object. Run a NCBI blast over it. Get the result. Run getbesthitsfromblast from the results. Run a Clustal from the results. Create tree from the Clustal results. Visualize the intermediate results. Definitions: o Blast: This tool finds regions of local similarity between sequences. o getbesthitsfromblast: Get the best hits from a Blast report. o Clustal: is a general purpose multiple sequence alignment program for DNA or proteins. This exercise can be followed by video tutorials, go to This exercise must be done in the INAB repository to ensure the availability of the Web Services. Click on change repository and change it if you are in other repository. Supplementary Material 13

14 Using Magallanes and the Object editor to standardize a sequence in plain text: More information about Magallanes and Object editor can be found at: and Click on the Search button 2. Write seqnce amnoacids in the text box 3. Magallanes will ask you if you mean sequence aminoacids 4. Click on Yes 5. Results will be displayed 6. Double click on Aminoacid Sequence data type. 4 6 Supplementary Material 14

15 Creating AminoacidSequence Object A video-tutorial of this section can be found at Search CTLA4_HUMAN.txt file in the file system area. 2. Drop it into the sequence String parameter area. 3. A dialog window will appear, click on Drop de content here option. 4. Write CTLA4_HUMAN.Object in the file name box. 5. Click on create Object. 6. The new file will appear in the file system area. 7. Search the new file in the File System area. 8. Right click on the file. 9. Click on Exact Compatible tool > runncbiblastp1 6 Supplementary Material 15

16 Since the tool execution parameter interface has been opened using the File System Area the parameter sequence (AminoacidSequence) is filled. 10. Write the output file name for the Blast report: Blast-result.xml 11. Click on Launch Service When the execution finishes, the tab will inform the user and the results will be available in the File System Area. (Blast-result.xml). In the case of errors, the tab will turn red and an error message will be displayed to inform the user about what has happened in the execution. result Supplementary Material 16

17 Once the result has been stored in the file system, the user can launch a compatible web service or use a compatible viewer if exists. Now we will execute the web service getbesthitsfromblast. 12. Look for the Blast-result.xml file in the file system area 13. Right-click on the file and select exact compatible tool > getbesthitsfromblast. 14. When the parameter window finish opening, write as output file name: best-blastresult.xml 15. Click on launch service Service execution finishes and the result file will be available in the file system area. 13 To visualize the result click on the mouse right button and select XML Viewer. This result is a collection of sequence database-identifiers. Supplementary Material 17

18 Let s find the Aminoacid Sequences from those identifiers: Open Magallanes and search Sequence Aminoacid Collection 18. Double click on getaminoacidsequencecollection (the description fits with the desired action) 19. The tool parameter window will open. Drag and drop the bestblast-result.xml result file. 20. Write the output file name best-blast-sequences.xml. 21. Click on Launch service When the service finishes you will get the best-blast-sequences.xml file in the File System Area. Now we will perform a multiple sequence alignment with the results: Search Clustal in the input box on the top of Control Area. 23. Matching tools will highlight 24. Double click on runclustalwfast 24 Supplementary Material 18

19 25. Drag the result file best-blast-sequences.xml to the sequences input box 26. Write the output file name clustal-result.xml 27. Click on Launch service and wait.. (this one takes some minutes) 28. Visualize it with JalView: Right click on the clustal-result.xml and select JalView viewer Supplementary Material 19

20 Now we are going to create a Newick tree from the results of the multiple alignment using a Web Service and a viewer: 29. Right click on the clustal-result.xml file and Exact compatible tool 30. Search for runcreatetreefromclustal tool. 31. Write NewickTree-file.xml in the output parameter. 32. Launch the tool 33. Visualize the result right clicking on the result file and selecting HyperTreeViewer Supplementary Material 20

21 Saving the exercise as a workflow: 34. Click on the LOG icon: executed tasks will be displayed 35. Drag and drop the executed tasks (in descendant order) into the blank layout. 36. Click on the export button and select a location to save the SCUFL file. Supplementary Material 21

22 Using Web Services - different protocols and repositories Web Services interoperability is managed by jorca based in a common data model e.g. BioMoby- for the inputs and outputs adopted by the Web Services providers. In the same line, inter-repository compatibility can only be guaranteed by adopting a common data model (e.g. the ACGT and the INB repositories are compatible with BioMoby Central repository since they both use the same data types definitions. However, it could happen that two different Web Services have different definitions of the same type of data (e.g. Moby and WSDL Web Services). For those cases, to solve the I/O matching it is necessary an intermediate process to transform the format of data. In jorca, this work is done helped by the BioData-SF software API ( In this exercise: o o o o o We will use the WABI Web Services to run a Blast over SWISS Prot Database. Wait for the result. Change to MOBY Central repository. Create a BioMoby Object with the WABI result and the automatic conversion method. Run a tool to predict a secondary structure and accessibility using PHD (neural networks methods) The first step is changing the repository to WABI. 1. Click on the change repository button. 2. Search for WABI repository and click on Change. 3. Search for searchsimple service in the Tool tree and look for the searchsimple operation in Blast. Then, double click on the operation Supplementary Material 22

23 5 4c 4. Fill the parameters: a. blastp in program b. SWISS in database c. Drag and drop the Q9M4M3_WHEAT.fas file into query and click in the Drop the content option. 5. Write BLAST-WABI.txt in the output box. 6. Click on Launch service (this process may take some minutes) 9a When the execution finishes the result will be available in the File System Area. 14 9b The result is a common BLAST-text file with some headers. 8. Change to MOBY Canada repository. 9. Use the advanced search button to open Magallanes Search GUI a. Write PHD in the search box. b. Look for the PHD service that uses a Blast-Text: runphdfromblasttext and double click on it. 10. The service parameter composition window will open. 11. Click on the Create new object button which is next to the Blast-text parameter. The datatype composition 13 panel will open. 12. Drag and drop the BLAST-WABI.txt file from the File System Area to the Create data type panel. c. The heuristic system detects that the data is not in plain format and offers to the user to convert it into plain format. d. Click on Try to convert it button, the plain Blast-text will be dropped into the text area box. 13. Write Blast-text.MOBY.xml in the file name input. Supplementary Material 23

24 14. Click on create object. 16 The new object will appear in the File System Area and the parameter from the web service will be automatically filled. 15. Click on Launch service and wait for the result. 16. When the tool finishes the execution, the result will appear in the file system. 17. Click on the result and XML Viewer to visualize the result. Supplementary Material 24

25 Using Web Services in a GRID environment With using Web Services in a GRID environment we are referring to invoking Web Services registered in a secure GRID repository (Globus based environment). In these environments, it is necessary to have a user, a password and credentials to execute Web Services. Ask vickymr@uma.es/maxgarcia@uma.es for this information if you are interested. Follow this full exercise at: 1. Click on the change repository button. 2. Search for ACGT Testing Repository repository and click on Change. 2 Once the repository is loaded: 3. Drop pre-processing web service from favourites to the main area 4. Drag and drop the GE-demo.dat file to the Gene Expression Matrix parameter 3 input box Use these parameters: logarithm base : 2 Standard Deviation filter : 0.5 Centring : normalize 6. Click on Launch service button 7. If the user hasn t configured the grid credentials a pop-up dialog will be shown, fill the dialog with your credential s information Once the information is complete, the tool will be launched. Supplementary Material 25

26 9. When the execution finishes, the tab will inform the user about the status, if the tool has executed successfully, the tab will turn green and the results will be available in the file system area. 10. Double click on the results files to visualize them with the default viewer. Supplementary Material 26

27 Executing R-BioMoby Web Services R is a wide-used software environment for statistical computing and graphics in the scientific community. Although jorca main objective is to facilitate the use of Web Services, R is not a Web Service platform (it is a script console for statistical research) but there are different implementations to use its functionalities as Web Services( Approaches to R as web and analytic service by Nianhua Li, Martin T. Morgan, Seth Falcon, Robert Gentleman available at pdf), therefore users could work with R using jorca if they had those Web Services available. In our group we have developed -as a proof of concept- some BioMoby Web Services that are able to call to R scripts. We could discern two different approaches: 1. Create an R Web Service that contains an R-script to perform an operation, plot etc. 2. Create an R Web service that receives the input file and the script to execute (this can produce security holes taking in account that several commands that affect the local file system can be executed). We have decided, for simplicity and security to use the first approach. The Web Service is implemented using a Perl script (front-end that transforms the data between BioMoby format and R-format) that takes a file with the R script and the inputs, this script is in charge of wrapping the inputs and the outputs and calling to the R console with the script and the data as inputs. The implemented R Web Service must be registered in a repository to be used by jorca or any Web Service client. The data management in R is usually performed using plain text, for that reason we have created two new data types that encapsulate this content. Next exercise provides guidelines to execute the implementation of R BioMoby Web Service. The BioMoby R-Web Services will be used to get a Box plot using a plain-text data file We are going to switch to the INB Testing repository (where the R-WS is registered).click on the change repository button. 2. Search for INB Testing repository and click on Change. 3. Use the R-data.txt file available in the files provided for the guided-exercise. This file contains a set of volatile measurements in plants; in text tab separated value format. Each line in this file (except the first that are the headers) contains a vector of values <Crop/Replication/Experiment/Variables>: Crop Replication/Number Experiment Var 1 Var 2 Var 3 0 ESCPIM001 1 EELM-2005-RILs NoData NoData NoData Supplementary Material 27

28 4. Search for BoxPlot using the quick search box; and double click in the result. 5. The tool parameter window will open. Click on the create object button. 6. The datatype composition window will open. Drag the Rdata.txt file and drop it into the input box An alert dialog-box will open. Click on the Drop the content here button. 8. Write the object name: RObject.xml 5 9. Click on create object, the new data object will be filled in the Web Service s parameter. 10. Launch the RBoxPlot web service. 11. When the Service finishes the execution, the result will be ready in the File System Area. Right click on the result and click on postscript viewer. (Only available for Windows OS) This exercise can be repeated using the same data object with the Web Services: RCorrelationPlot, RHistogramPlot. 7 Supplementary Material 28

29 5. Guided exercises on data standardization Using the data conversion wizard Use the file system area to move into the raw folder, you will find several files with data in plain format. 1. Simple txt sequence (1seq.txt) 1. Right click on the file and select the Convert to.. Submenu. 2. Select a format. 3. Write the output file name. 4. Double click on the new file to visualize the results Three Sequences in the same (3seqs.fas) 1. Right click on the file and select the Convert to.. Submenu 2. Click on Data conversion wizard 3. Select a sequence from the detected and click on Next 4. Select the format and write the output file name, click on finish 5. Double click on the new file to visualize the results Supplementary Material 29

30 3. PDB-Text (PDB-raw.txt) 1. Right click on the file and select Convert to.. Submenu 2. Click on Data conversion wizzard 3. Select a sequence from the detected or the PDB text object and click on Next 4. Select the format and write the output file name, click on finish 5. Double click on the new file to visualize the results Supplementary Material 30

31 On the fly data standardization On the fly data standardization is performed when a raw data file is dragged into a input that needs a XML-standardized object. 1. Search the runncbiblastp1 web service. 2. Double click on the tool. 3. Drag and drop the file 3seq.fas from the File System area to the input box. 4. Choose Try to standarize 5. If any compatible sequence is found it will be displayed, choose one of the results. 6. Click on launch Tool and wait for the result Supplementary Material 31

32 6. Comparison of Web Services clients We will perform a one by one comparison of Web Services clients by the features analyzed in the manuscript. Analyzed clients: GBrowse (Wilkinson M. 2006) MOWServ (Navas et al., 2003) Dashboard( ) Seahawk (MK. Gordon P. et al, 2007) BioCatalogue Client ( Taverna (Oinn et al. 2004) Remora (Carrere S. et al, 2005) jorca Multi-repository integration GBrowse MOWServ Dashboard Seahawk It accesses and uses the BioMoby Central Repository. It exploits the INB extended BioMoby repository, it can access to any local instance of the INB BioMoby extended repository. It is able to access and register services in any BioMoby repository using its endpoint URL. It accesses and uses the BioMoby Central Repository. But It is able to access and register services in any BioMoby repository using its endpoint URL. BioCatalogue It is a collaborative centralized repository, it accesses its own repository but service providers can register their services using their WSDL description. Taverna Remora jorca Discovery methods GBrowse MOWServ It is able to read and access any BioMoby repositories, WSDL Web Services, BioMART repository and SoapLab Web Services, also Beanshells and local R scripts. It is able to access to any BioMoby repository using its endpoint URl. By default it is able to access INB Repository, INRA repository, MOBYCentral repository and OpenCentral Repository. It is able to retrieve the tools and data-types from: BioMoby Repositories (Central, Development and INRA), The national institute of Bioinformatics of Spain, the ACGT repository (which requires executing in secure environments) and access to WABI Japanese and EBI WSDL Web Services. It uses a data-driven Web Service discovery system. When starting the user must provide a starting point conformed by a Namespace (abbreviation for a database), and an ID (identifier). Once selected it searches for the appropriate bioinformatics services services that consume the initialized data-type. The user can execute it and can obtain the compatible Web Services with the result. It provides two alternatives for service discovery: a. the BioMoby taxonomies for services and data-types are represented in a tree that the user can browse and search, but limited to full-string matching and b. Services can be located based on the input data matching data-types semantic descriptions. Dashboard It provides a syntactic search over the different tool trees that highlight the tools that match with the searched tool name. It is able to use regular expressions. Seahawk It attempts to match the user s biological data type with the declared input and output data-types of Web Services, and then present the user with the next possible steps. BioCatalogue It provides a keyword based search and allows filtering the results by some metadata like service type, provider, country, etc. Taverna In the 1.7 version the search is limited to highlight in the different scavengers the service names that syntactically match a word without filtering. In the 2.0 version the services can be ordered by several parameters (authority, URI, type, etc.) the search results are filtered and has also a plug-in for FETA that allows the user to discover services based on name or additional concept-based constraints the user can select from a list (task performed, data resource used, method used, I/O data-types, etc.), the concepts used in the search are taken from the mygrid ontology. Remora It has a search method to look for a keyword in the service description or name and input types; it also displays available Web Services based on the currently produced data type in the workflow construction. Supplementary Material 32

33 jorca jorca provides different alternatives for discovering tools: a) Syntactic text-search engines and filtering mechanisms highlight identified Web Services, data-types and namespaces using taxonomy trees, plain lists, by authorities, filter the trees by offline/online etc; b) Plug-ins connect external applications to incorporate new search strategies (Magallanes); c)input/output matching of services, i.e. What services are able to process my sequence? ; d) Advanced search mechanisms for pipelining several tools to solve a complex query; i.e. How can I obtain a phylogeny for the gene I have? ; and e)giving the user the ability to build his/her own tool repository on the fly using a favourites mechanism. Service composition GBrowse MOWServ Dashboard Seahawk Given an ID and a namespace as the starting point it allows pipelining Web Services matching input/output, the extra-input parameters are prompted to the user using web forms. It uses automatic interface builder based on the service s metadata descriptors and present them in a web form to be filled by the final user. MOWServ is also able to chain services by matching output with input. Dashboard uses automatic interface builders based on the service s metadata descriptors. The data must be included in an input text element; it doesn t allow loading a data file. An advantage is how it manages the objects collections. Seahawk addresses the problem of data type definition with a data-driven approach to execute BioMoby services. In the pasted text or in a link, Seahawk recognizes the selected text as strings or sequences. It automatically converts to a BioMoby object and launches a compatible web service with that data type. BioCatalogue In the current version of BioCatalogue s we didn t find the option of executing Web Services. Taverna Service composition is made as a test of the workflow, the user must create an input object and complete the object in text-plain, not only the XML-object also the protocol message headers. It is not very intuitive for final users, good for service and workflow developers. Remora Once the workflow has been defined it allows adding input data using web forms automatically built using the service metadata. jorca Different methods have been implemented in jorca to complete the needed input data and enable tools execution. The aim is to close the gap between traditional data and structured data formats, giving support to a broad range of user skills: a) Automatic and generic interface creation completed with default values b)heuristic methods to find the appropriate input files among the user data; and c)on the fly data standardization Service invocation GBrowse MOWServ Dashboard Seahawk GBrowse executes BioMoby Web Services. Once the service pipelining and execution has finished, GBrowse allows exporting the pipeline as a SCUFL workflow. MOWServ executes BioMoby Web Services and SCUFL Workflows. It only executes BioMoby Web Services. It only executes BioMoby Web Services using a data-driven approach. BioCatalogue BioCatalogue current version is not able to execute Web Services. Taverna Taverna is able to execute WSDL Web Services, SOAPLab Web Services, BioMoby Web Services but the user must take in account the protocol, the service-protocol headers must be included in the inputs and outputs of the workflow or services, this feature is hard to use by biologist or technical people without expertise in Web Services. Remora When the workflow has been submitted and the input data had been included, it allows executing it. Remora asks for an address and a name for a workflow and the user when the process has finished. It only executes BioMoby Web Services. jorca The current jorca version has workers that interact with: BioMoby asynchronous and synchronous services, WSDL defined Web Services and the invocation of tools inside secure environments, the ACGT architecture supposes the integration of different tools and data sources, where some of them require a previous process of Anonymization due to privacy policies. Once this data is available, the final user can analyze it inside the Grid environment. These requirements features a high grade of security, which is covered making use of different tools like GAS (Gridge Authorization System), were all user credentials are supported. New workers components can be easily plugged into jorca to invoke Web Services with different protocols such as command line programs, R-scripts ( and SCUFL ( which will be included soon. jorca logs all the user activity locally, and lets user re-launch tasks and create personalized pipelines from a set of selected operations. These pipelines can be created as simple scripts in a standard language (SCUFL) and used in specialized graphical workflow tools such as Taverna or Triana ( Supplementary Material 33

34 2.2.6 Usability GBrowse MOWServ Dashboard It is the first Moby client ever written; very simplistic and limited in power, but easy to understand and use for beginners. This browser uses BioMoby a data-driven Web Service discovery system. It s limited but good to start in the use of Web Services. MOWServ is a web client for service discovery and execution, object creation with user authentication system. It is a versatile client that access and extend BioMoby repositories. It has automatics interfaces generator that reduces the developing time and produces uniform service and data-types access mechanisms. It interface is not very user friendly for final users. Dashboard is a web service providers/developers oriented standalone client that allows the registration/test of Web Services and contain an interface to create and display MOBY-S objects.(but not saving them, only for testing issues). It does not allow invocating pipelined Web Services. It is a very good application for developers. Seahawk Seahawk is a standalone application that allows the execution of Moby Web Services through the data. It is easy to use but limited to the data-types defined in the rule heuristic systems. BioCatalogue BioCatalogue is a nice web application for browsing the different Web Services, browsing the catalogue is friendly and has several aids that allows filtering the results by keywords, service type etc. Although is not able to execute Web Services yet. Taverna Taverna is useful to develop workflows, it is able to read and execute different Web Services protocols but it lacks in usability. There is a full study about Taverna usability at This document states: Many of the problems in Taverna are related to the user interface, or program installation, and therefore could be addressed by motivated developers. The major, more fundamental issue is data input types/formatting/connecting, which were showstoppers for users in this study. Taverna is fundamentally data format agnostic. Given users reported willingness to re-use and modify analysis workflows, it may make more sense to provide end-users with existing workflows that accept and enforce semantically typed data. Remora This client is a workflow design and executing tool for BioMoby Web Services. It starts by selecting a datatype or searching a tool and it guides the user through the next possible Web Service in the pipeline matching input/output. It is not possible to export the generated workflow. The interface is clear and easy to use but it is limited to executing created workflows in their own system (not standardized). jorca The jorca interface aims to provide a comfortable environment for the integrated exploitation of webresources A clear focus of development was to create friendly features that appeal to a broad range of users Every user can customize the interface by building most used services and data-types in the Favorites section. Since most users are not familiar with Web Services, data-types and namespaces terminology, jorca can show or hide these information trees, or select among multiple displays. The user can also personalize jorca s look and feel by selecting one of the provided skins or creating their own. File system browsing, drag and drop and double-clicking icons are common and successful features of GUI clients in computing today. jorca reproduces these features in its graphical interface. The file system manager in jorca lets the user copy, delete, rename, etc., and organize folders hierarchically. These traditional features are enhanced with jorca s ability to identify the set of operations available for files: XML, HTML or text based visualizations, editing functionality, plus its dispatch system to invoke a service using the selected file as main input. The jorca package is completed with online updated help and documentation system, tutorials, training material and guided exercises. Next table will resume these features. Supplementary Material 34

35 GBrows MOWServ Multi repository Discovery method Service composition Service invocation BioMoby Central. Any INB-BioMoby extended repository. (one per instance). a) Keyword based (Namespace, DB, ID). b) Input/Output tool matching. a) Browsing trees. (Tools and data types). b) Syntactic text-search engines with highlight mechanisms. c)input/output tool matching. a) Compatible tools with a result. b) Automatic and generic input form. a) Compatible tools with a result. b) Automatic and generic input form. a) Executes BioMoby Web Services. b) Exports the execution to SCUFL workflow. a) Executes BioMoby Web Services. b) Executes SCUFL workflows. Dashboard Any BioMoby repository. a) Syntactic text-search engines and regular expressions with highlight mechanism. b) Browsing trees (authority, alphabetically, by inputs/outputs) a) Automatic and generic input form. Collections are easy to handle. a) Executes BioMoby Web Services. Remora Taverna BioCatalogue Seahawk Any BioMoby repository. Using a rules based system, data types are found and compatible tools are displayed. its own repository. Service providers can register their services using the WSDL description. Any BioMoby repository. WSDL Web Services. BioMART repository. SoapLab repository. Beanshells. local R scripts. a) Keyword based search. b) Metadata results filter (service type, provider, country, etc.). (version 2.0) a) Information organization (authority, URI, type, etc.). b) Syntactic text-search with filtering mechanisms. c) plug-in for FETA (matching by concept-based constraints). Any BioMoby repository. a) Keyword based search. B) Input/Output tool matching. In the pasted text or in a link, Seahawk recognizes the selected text as strings or sequences. It automatically converts to a BioMoby object and launches a compatible web service with that data type. No execution option in the current version. Service composition is made as a workflow test. The input object must be created, not only the XML-object also the protocol message headers. Not very intuitive for final users, good for developers. a) Automatic and generic input form. a) Executes BioMoby Web Services. No execution option in the current version. a) Executes WSDL Web Services. b) Executes SOAPLab Web Services. c) Executes BioMoby Web Services. d) Executes Beanshells. e) Executes local R scripts. a) Executes BioMoby Web Services. jorca Any BioMoby repository. Any INB-BioMoby extended repository. WSDL Web Services. ACGT repositories (Grid secure environment) a) Syntactic text-search with filtering mechanisms and highlight results. b) Plug-ins for external engines (Magallanes) c)input/output tool matching d) Favourites mechanism. e) Different information displays. a) Automatic and generic input form. b) Heuristic methods to find the appropriate input files among the user data. c) On the fly data standardization. d) Input/output tool matching. a) Executes BioMoby asynchronous and synchronous Web Services. b) Executes WSDL-SOAP defined Web Services. c) Executes tools in secure environments. d) New workers components can be plugged (command line programs, R-scripts and SCUFL). e) Logs all the user activity, tasks can be re-launched. f) Allows creating personalized pipelines in SCUFL. Supplementary Material 35

36 7. Integration using the Modular API In jorca, inter-repositories integration is done helped by the Modular API framework -mapi- (Ramirez, S. et al International Conference on Telecommunications and Multimedia. A flexible framework for the design of knowledge-discovery clients ). mapi is a flexible and expandable framework that allows fast development of clients that access distributed computational resources. The framework is organized in several modules, each in charge of a specific task, such as tool metadata maintenance and discovery, file browsing and user administration. By splitting the functionality in modules, is possible to provide a flexible framework where modules can be combined and adapted to project-specific requirements. For example, jorca doesn t use Users-groups modules or the methods to save information in repositories. The modules developed are organized following a hierarchy of dependencies, as can be seen in the next figure. The most basic module allows the definition of available data types in the system. Two main branches of dependencies correspond to Tools (software components that consume and/or produce data) and instances of a particular data-type (Data). The enactor combine these two branches by sending the data to the tools for execution using the information stored in the mirror module. The modules related to access-rights (users and groups) can be combined with every module providing authentication and access rights for resources. This figure shows dependencies between modules of the framework. User-Group modules which are not shown for simplicity interact with all the other modules providing authentication and access rights. Each module is composed by several layers. The layers can be combined in several ways, allowing clients to work independently of metadata location (local or remote) and the access protocol. Supplementary Material 36

37 Clients connect to modules through the following layers: Application interface layer: This is the highest level programming interface. jorca access and use this interface. Local/remote access layer: Depending on configuration, the application interface layer accesses either a local implementation (if metadata is available locally) or a remote access layer (if metadata is available on a remote server). These access layers provide a uniform view of metadata by handling multiple repositories in a common way, independently of the access protocol and location of the metadata. Extended information (user guide, submitted manuscripts, examples, descriptions, libraries etc.) can be found at: The virtual representation is performed by mapi which provides to jorca a uniform view of the different repositories and protocols. Thanks to mapi, jorca retrieves the tools, data type, namespaces and functional categories from different repositories, manages all that information using its own modules and combines and displays it to the final user through friendly interfaces. jorca also uses the Execution Module to execute the different Web Services. Supplementary Material 37

38 9. Model View Controller Pattern In the Model View Controller Pattern ( the model represents the information of the application and the business rules used to manipulate the data, the view corresponds to elements of the user interface such as text, checkbox items, and so forth, and the controller manages details involving the communication to the model of user actions such as keystrokes and mouse movements. Supplementary Material 38

39 10. Ongoing work Reading and executing SCUFL workflows Heuristics: Create more rules for automatic data conversion. Improve performance with large files (right now is limited to small ones) Advanced execution of WSDL services. Supplementary Material 39

40 11. Feedback Any suggestions, bugs or feedback vickymr at uma.es ortrelles at uma.es Keep updated at: Full help at: Video tutorials at: Supplementary Material 40