I. Wassink

Performing statistical analyses on quantitative data in Taverna workflows: An example using R and maxdBrowse to identify differentially-expressed genes from microarray data

BackgroundThere has been a dramatic increase in the amount of quantitative data derived from the measurement of changes at different levels of biological complexity during the post-genomic era. However, there are a number of issues associated with the use of computational tools employed for the analysis of such data. For example, computational tools such as R and MATLAB require prior knowledge of their programming languages in order to implement statistical analyses on data. Combining two or more tools in an analysis may also be problematic since data may have to be manually copied and pasted between separate user interfaces for each tool. Furthermore, this transfer of data may require a reconciliation step in order for there to be interoperability between computational tools.ResultsDevelopments in the Taverna workflow system have enabled pipelines to be constructed and enacted for generic and ad hoc analyses of quantitative data. Here, we present an example of such a workflow involving the statistical identification of differentially-expressed genes from microarray data followed by the annotation of their relationships to cellular processes. This workflow makes use of customised maxdBrowse web services, a system that allows Taverna to query and retrieve gene expression data from the maxdLoad2 microarray database. These data are then analysed by R to identify differentially-expressed genes using the Taverna RShell processor which has been developed for invoking this tool when it has been deployed as a service using the RServe library. In addition, the workflow uses Beanshell scripts to reconcile mismatches of data between services as well as to implement a form of user interaction for selecting subsets of microarray data for analysis as part of the workflow execution. A new plugin system in the Taverna software architecture is demonstrated by the use of renderers for displaying PDF files and CSV formatted data within the Taverna workbench.ConclusionTaverna can be used by data analysis experts as a generic tool for composing ad hoc analyses of quantitative data by combining the use of scripts written in the R programming language with tools exposed as services in workflows. When these workflows are shared with colleagues and the wider scientific community, they provide an approach for other scientists wanting to use tools such as R without having to learn the corresponding programming language to analyse their own data.

Analysing Scientific Workflows: Why Workflows Not Only Connect Web Services

Life science workflow systems are developed to help life scientists to conveniently connect various programs and web services. In practice however, much time is spent on data conversion, because web services provided by different organisations use different data formats. We have analysed all the Taverna workflows available at the my Experiment web site on December 11, 2008. Our analysis of the tasks in these workflows shows several noticeable aspects: their number ranges from 1 to 70 tasks per workflow; 18% of the workflows consist of a single task.Of the tasks used are 22% web services; local services, i.e. tasks executed by the workflow system itself, are very popular and cover 57% of tasks; tasks implemented by the workflow designer, scripting tasks, are is also used often (14%). Our analysis shows that over 30\% of tasks are related to data conversion.

Human-Centered Aspects

Humans have remarkable perceptual capabilities. These capabilities are heavily underestimated in current visualizations. Often, this is due to the lack of an in-depth user study to set the requirements for optimal visualizations. The designer does not understand what kind of information should be visualized, how it should be presented or what kind of interactions should be supported. The key elements of successful information visualization are the correct data using the best visualization technique and the best interaction techniques with respect to users. If one of these elements is ignored, people might interpret the data in the wrong way and thus might not understand the underlying information or a pattern. In order to design effective interactive visualizations, it is important to take into account the limitations of human perception, context of use, and the goals and activities that are to be performed to reach these goals. In order to obtain a usable application, developers have to pay attention to the user’s working environment and tasks; this focus-on-user idea is comprised in the human-centered concept. The next section discusses usability (the property of being usable) from the human-centered point of view. Usability has application in many areas, but our focus is on the human-centered approach to design of interactive systems, also called user-centered, in order to inform the reader on how to design visualizations according to human cognitive and perceptual abilities, specific to the context of use and goals of potential users. Then, the usability concept is explained in the “Usability in Human-Centered Design” section. The next Section “User Aims and Requirements” discusses how to define a user group, establish user goals and requirements. Finally, an overview of the different evaluation methods and current evaluation practices, including the practical issues of experiment design that can help to improve the effectiveness of visualizations is presented in the “Evaluation of Visualizations Environments” and “User Studies and a Science of Visualization” sections.

Applying a User-centred Approach to Interactive Visualization Design

Analysing users in their context of work and finding out how and why they use different information resources is essential to provide interactive visualisation systems that match their goals and needs. Designers should actively involve the intended users throughout the whole process. This chapter presents a user-centered approach for the design of interactive visualisation systems. We describe three phases of the iterative visualisation design process: the early envisioning phase, the global specification hase, and the detailed specification phase. The whole design cycle is repeated until some criterion of success is reached. We discuss different techniques for the analysis of users, their tasks and domain. Subsequently, the design of prototypes and evaluation methods in visualisation practice are presented. Finally, we discuss the practical challenges in design and evaluation of collaborative visualisation environments. Our own case studies and those of others are used throughout the whole chapter to illustrate various approaches.

Designing Workflows on the Fly Using e-BioFlow

Life scientists use workflow systems for service orchestration to design their computer based experiments. These workflow systems require life scientists to design complete workflows before they can be run. Traditional workflow systems not support the explorative research approach life scientists prefer. In life science, it often happens that few steps are known in advance. Even if these steps are known, connecting these tasks still remains difficult. We have extended the e-BioFlow workflow system with an ad-hoc editor to support on-the-fly workflow design. This ad-hoc editor enables an ad-hoc design of the workflow with no predetermined plan of the final workflow. Users can execute partial workflows and extend these workflows using intermediate results. The ad-hoc editor enables its users to explore data and tasks representing tools and web services, in order to debug the workflow and to optimise parameter settings. Furthermore, it guides its users to find and connect compatible tasks. The result is a new workflow editor that simplifies workflow design and that better fits the explorative research style life scientists prefer.

E-BioFlow: Different Perspectives on Scientific Workflows

We introduce a new type of workflow design system called e-BioFlow and illustrate it by means of a simple sequence alignment workflow. E-BioFlow, intended to model advanced scientific workflows, enables the user to model a workflow from three different but strongly coupled perspectives: the control flow perspective, the data flow perspective, and the resource perspective. All three perspectives are of equal importance, but workflow designers from different domains prefer different perspectives as entry points for their design, and a single workflow designer may prefer different perspectives in different stages of workflow design. Each perspective provides its own type of information, visualisation and support for validation. Combining these three perspectives in a single application provides a new and flexible way of modelling workflows.

Bringing hollywood to the driving school: dynamic scenario generation in simulations and games

In this paper we discuss a framework for simulation software called the movie metaphor. It is applied to the Dutch Driving Simulator for dynamic control of traffic scenarios. This framework resolves software complexity by the use of agent protocols inspired by the way of working on a movie set. It defines clear responsibilities for the agents so that the system is extensible, maintainable and easy to understand. The framework is a software pattern for multiagent systems especially suitable for simulation software and games.

Using R in Taverna: RShell v1.2

BackgroundR is the statistical language commonly used by many life scientists in (omics) data analysis. At the same time, these complex analyses benefit from a workflow approach, such as used by the open source workflow management system Taverna. However, Taverna had limited support for R, because it supported just a few data types and only a single output. Also, there was no support for graphical output and persistent sessions. Altogether this made using R in Taverna impractical.FindingsWe have developed an R plugin for Taverna: RShell, which provides R functionality within workflows designed in Taverna. In order to fully support the R language, our RShell plugin directly uses the R interpreter. The RShell plugin consists of a Taverna processor for R scripts and an RShell Session Manager that communicates with the R server. We made the RShell processor highly configurable allowing the user to define multiple inputs and outputs. Also, various data types are supported, such as strings, numeric data and images. To limit data transport between multiple RShell processors, the RShell plugin also supports persistent sessions. Here, we will describe the architecture of RShell and the new features that are introduced in version 1.2, i.e.: i) Support for R up to and including R version 2.9; ii) Support for persistent sessions to limit data transfer; iii) Support for vector graphics output through PDF; iv)Syntax highlighting of the R code; v) Improved usability through fewer port types.Our new RShell processor is backwards compatible with workflows that use older versions of the RShell processor. We demonstrate the value of the RShell processor by a use-case workflow that maps oligonucleotide probes designed with DNA sequence information from Vega onto the Ensembl genome assembly.ConclusionOur RShell plugin enables Taverna users to employ R scripts within their workflows in a highly configurable way.

e-BioFlow: improving practical use of workflow systems in bioinformatics

Workflow management systems (WfMSs) are useful tools for bioinformaticians. As experiences with using WfMSs accumulate, shortcomings of current systems become apparent. In this paper, we focus on practical issues that hinder WfMS users and that arise in the design and execution of workflows, and in access of web services. We present e-BioFlow, a workflow engine that demonstrates in which way a number of these problems can be solved. e-BioFlow offers an improved user interface, can deal with large data volumes, stores all provenance, and has a powerful provenance browser. e-BioFlow also offers the possibility to design and run workflows step by step, allowing its users an explorative research style.