Why use workflow? In recent years, explosive amounts of biological information has been obtained and deposited in various databases.

Transcription

1 Workflow technology It s a generic mechanism to integrate diverse types of available resources (databases, servers, software applications and different services). It facilitates knowledge exchange within traditionally divergent fields such as molecular biology, clinical research, computational science, physics, chemistry and statistics. 1

2 Workflow tecnology Researchers can easily incorporate and access diverse, distributed tools and data to develop their own research protocols for scientific analysis. 2

3 Applications Application of workflow technology has been reported in areas like drug discovery, genomics, large-scale gene expression analysis, proteomics, and system biology. 3

4 Why use workflow? In recent years, explosive amounts of biological information has been obtained and deposited in various databases. Workflow systems could become crucial for enabling scientists to deal with this data explosion. A workflow system is highly flexible and can accommodate any changes or updates whenever new or modified data and corresponding analytical tool become available. 4

5 How workflow works Workflow environment allows life science researchers to perform the integration themselves without involving any programming. Workflow system allows the construction of complex in silico experiments in the form of workflows and data pipelines. 5

6 Data Pipelines Data pipelining is a relatively simple concept. Any computational component or node has data inputs and data outputs. Data pipelining views these nodes as being connected together by pipes through which data flows. In a workflow controlled data pipeline, as the data flows, it is transformed and raw data is analyzed to become information and the collected information gives rise to knowledge. 6

7 Workflow description A workflow system (Hollingsworth, 1995) is a holistic unit that defines, manages, and executes workflow processes aided by software. The order of execution is defined by a computer representation of the workflow process logic. 7

8 Workflow description Internally, a workflow system uses a Workflow Language or Meta-Languages for process specification (Michael and Jörg, 1999) to define the workflow process logic, to be executed by workflow execution engine or workflow controller 8

9 Workflow description (2) Visual representation of the workflow process logic is generally carried out using a Graphical User Interface (GUI) where different types of nodes (data transformation point) or software components are available for connection through edges or pipes that define the workflow process 9

10 Workflow component The anatomy of a workflow node or component is basically defined by three parameters: input metadata transformation rules, algorithms or user parameters, output metadata. 10

11 Anatomy of a workflow node The component properties are best described by the input metadata, output metadata and user defined parameters or transformation rules. The inputs ports can be constrained to only accept data of a specific type such as those provided by another component 11

12 Node linking Nodes can be plugged together only if the output of one, previous (set of) node(s) represents the mandatory input requirements of the following node. Thus, the essential description of a node actually comprises only in and output that are described fully in terms of data types and their semantics. 12

13 A general workflow based framework for life sciences domain 13

14 A general workflow based framework for life sciences domain A general workflow has three different layers: clients layer, component and enactment layer, and database layer 14

15 Clients layer The client has a Graphical User Interface for the creation of workflows along with process definition service. The user can create workflows using any combination of the available tools, services or databases in workflow system by dragging/dropping and linking graphical icons. Process definition services use Meta-Languages for Workflow and Process Modeling. 15

16 Component and enactment layer This layer has component services and enactment services and it can interact with database layer. Component services provide different software components, tools and services (grid An enactment service consists of one or more workflow engines in order to manage and execute workflow instances. Workflow engine or workflow controller is responsible for the part or all of the runtime control environment within an enactment service. and web services). 16

17 Database layer Database layer consist of diverse range of data sources (remote as well as local). 17

18 Commercial Workflow Systems SciTegic s Pipeline Pilot chemistry (compound library acquisition, combinatorial library design, molecular property calculators, filters, and manipulators), ADME/Tox, Decision Trees, Modeling, R Statistics, Reporting, Sequence Analysis, BioMining, Text Analytics and Integration Collection (flexible mechanisms to link external applications and databases) 18

19 Commercial Workflow Systems InforSense KDE environment and its specialized extension BioSense High performance bioinformatics solutions ranging from sequence analysis to microarray informatics and remote database annotation ChemSense High range of chemoinformatics solutions ranging from the analysis and visualization of chemical libraries to the development of combinatorial chemistry libraries, and includes a wide range of QSAR, ADME-Tox prediction, molecular modeling and evaluation methods 19

20 Commercial Workflow Systems InforSense KDE and SciTegic s Pipeline Pilot are state of the art workflow systems helping faster and efficient research in life sciences domain. Due to the high costs involved, these are still not accessible to academics a major contributor to scientific research. Other commercial workflow system to be mentioned are INCOGEN VIBE, BioLog s BioLib, Science Factory s übertool. 20

21 Public Domain Workflow Systems There are dozens of open-source workflow systems in the life sciences domain, many of which were initially developed as specific and often small-scale projects. 21

22 Public Domain Workflow Systems Opensource workflow systems represent a major source of advantage for the academia not just because they are free of license charges, but also because the open-source workflow systems are based on community models of development in which people from diverse background actively contribute to the application. 22

23 Public Domain Workflow Systems KNIME (University of Konstanz) It is based on the Eclipse platform and, it provides an excellent data mining platform for drug discovery informatics, bioinformatics and chemistry research Taverna It addresses problems beyond the capabilities of the present system to improve many areas including Data flow centric model, Scalability and Data streaming. 23

24 Public Domain Workflow Systems my Grid It is developing semantically enabled grid middleware for supporting bioinformatics and drug discovery applications and, is regarded as the most powerful workflow system in public domain. Different application tools built on mygrid cover fields like Genome and Proteome Annotation (e-protein), Integrative Systems Biology (myib), Integration of Biological Data (PlaNet, EMBRACE), Proteomics (ISPIDER), Medical and Health Informatics (PsyGrid, MIAS-Grid), Chemoinformatics (CDK- Taverna, Chimatica) andtext mining (Epitheliome). 24

25 Taverna Taverna project aims to provide a language and software tools to facilitate easy use of workflow and distributed compute technology within the escience community Taverna is available freely under the terms of the GNU Lesser General Public License (LGPL) and it is possible to download at the site Taverna is written in Java 5, and runs on a selection of platforms. 25

26 Taverna Taverna allows a biologist or bioinformatician with limited computing background and limited technical resources and support to construct highly complex analyses over public and private data and computational resources, all from a standard PC, UNIX box or Apple computer. 26

27 Taverna The screenshot below shows the workbench in action running an example workflow. 27

28 Workflow example 28

29 Specialized Softwares for Integration with Life Sciences Workflow Systems Within the last few years a large number of tools and softwares dealing with different computational problems related to life sciences have been constantly developed Incorporating third party or new tools into existing frameworks needs a flexible, modular and customizable workflow framework A plug-in based architecture (like Eclipse) is a better option for an ideal workflow system. The workflow execution engine or controller communicates with the integrated components plug-ins using workflow language or XML. 29

30 Desired Traits and Future Trends Workflow systems can be data-intensive, computationintensive, analysis intensive, visualization-intensive (e.g., visualization pipeline systems such as AVS, OpenDX, and SCIRun), process-intensive or a combination of one or more traits. 30

31 Desired Traits and Future Trends Workflow systems can be data-intensive, computationintensive, analysis intensive, visualizationintensive (e.g., visualization pipeline systems such as AVS, OpenDX, and SCIRun), process-intensive or a combination of one or more traits. There are many common desired traits (Ludäscher et al., 2005) of workflow systems like seamless access to resources and services, service composition and reuse, scalability, detached execution, reliability and faulttolerance, semantic binding, process provenance and data provenance. 31

32 Desired Traits and Future Trends Present workflow systems in life sciences need to integrate several resources like Semantic web technology (Shadbolt et al., 2006), grids services and web services. Web and grid services provide access to distributed resources, while workflow techniques enable the integration of these resources to perform in silico experiments. 32

33 Conclusion We described workflow based framework to speed up the life sciences research. Largescale data analysis needs flexible workflow based integration of different software tools and application from diverse domains which can provide in silico experimental design through visual programming and execution on grids. 33

34 Conclusion Workflow systems in life sciences are still evolving and the final goal is a distributed and a ubiquitous environment which can integrate web services, semantic web, grids, domain specific tools and ontologies. 34