The Enzyme Portal: a case study in applying user-centred design methods in bioinformatics

Abstract

User-centred design (UCD) is a type of user interface design in which the needs and desires of users are taken into account at each stage of the design process for a service or product; often for software applications and websites. Its goal is to facilitate the design of software that is both useful and easy to use. To achieve this, you must characterise users' requirements, design suitable interactions to meet their needs, and test your designs using prototypes and real life scenarios.For bioinformatics, there is little practical information available regarding how to carry out UCD in practice. To address this we describe a complete, multi-stage UCD process used for creating a new bioinformatics resource for integrating enzyme information, called the Enzyme Portal (http://www.ebi.ac.uk/enzymeportal). This freely-available service mines and displays data about proteins with enzymatic activity from public repositories via a single search, and includes biochemical reactions, biological pathways, small molecule chemistry, disease information, 3D protein structures and relevant scientific literature.We employed several UCD techniques, including: persona development, interviews, 'canvas sort' card sorting, user workflows, usability testing and others. Our hope is that this case study will motivate the reader to apply similar UCD approaches to their own software design for bioinformatics. Indeed, we found the benefits included more effective decision-making for design ideas and technologies; enhanced team-working and communication; cost effectiveness; and ultimately a service that more closely meets the needs of our target audience.

Figure 1

Outline of the user-centred design process used during the project. The boxes indicate the approximate time taken to complete each step and the dotted lines indicate the scope of influence they had on the project. Usability testing has continued right through to the interactive prototype. Software implementation started once the first version of the specification was available. Note: ‘personae’ are user profiles; see details in the main article.

Figure 2

Personae were developed by the stakeholders and refined by user interviews. Two examples of the personae we created are (a) ‘Eunice’, and (b) ‘Debra’. Both represented bench scientists, with Eunice working in enzymology, and Debra in drug discovery. These personae, and three others shown in the Additional files 3 (‘Eric’ and and ‘Dean’) and 4 (‘Brenda’), reflected the expected desires and behaviours of potential users of the Enzyme Portal website. We described them in terms of name, role and motivations for using an enzyme resource.

Figure 3

Excerpt of the workflow analysis for the Enzyme Portal. The workflow is a large and comprehensive model of how the user journeys for various personae interact with the Enzyme Portal, and serves to highlight commonalities in behaviour across personae. The thought bubbles indicate the motivation for the user to come to the Enzyme Portal, and the arrows indicate the flow through the layers of the site. The boxes represent individual webpages and list the data items expected to be displayed at each stage of the journey. The complete workflow is available in the Additional file 5.

Figure 4

Photograph of the materials used in the canvas sort activity with user workshop participants. A novel ‘canvas sort’ approach was used in workshops with users to elicit information architecture, and to prioritise enzyme-related data items.

Figure 5

Photograph of a completed artefact generated in the canvas sort activity by user workshop participants. These artefacts were useful for prioritisation of data items and functionalities when planning the structure of the Enzyme Portal. There were five teams and each team produced 3 or 4 of these in the workshops.

Figure 6

Concept diagram of the information architecture applied in the Enzyme Portal. The data objects are grouped and each category has a number of data items valuable to the user. Categories and data objects are then interconnected. Concepts that did not fit into this model were disease and literature, which span across all the data objects and categories.

Figure 7

The paper prototyping method was used for usability testing of initial designs for the Enzyme Portal. The user was presented with a paper wireframe of each screen interface and could interact with it using a pencil instead of a mouse and keyboard. Each screen would then change according to the interaction. The wireframe interface designs were created using Balsamiq software (http://balsamiq.com).

Figure 8

The first version of the search results page using an interactive prototype. In this version not all the protein structures were displayed and also the species were located on the right hand side of each entry found.

Figure 9

A later version of the search results page after improvements highlighted during interactive usability testing. In this version the species are located below the synonyms, and protein structures is displayed for an entry irrespective of the default species.