Let the shoemaker make the shoes - An abstraction layer is needed between bioinformatic analysis, tools, data, and equipment:An agenda for the next 5 years

摘要

Bioinformatic tool development is being driven by individual efforts which while extending the boundaries of what is possible, are constrained by the framework in which the tools are being defined. This is resulting in a slower development process, aswell as tools that operate independently of other tools, and suffer inconsistent interfaces in terminology, layout, level of standards compatibility, stability, etc. The utility of these tools could be increased with better planning and development during this growth stage of bioinformatics tool development but this requires the adoption of a 'grander plan' of how the architecture of bioinformatics should be laid out.There are a number of themes to this discourse.It is economically attractive to allowspecialists to focus on their respective scientific specialities rather than on building a computing framework and the associated tools necessary to pursue their respective specialities.There is a body of knowledge in how to best utilise individual tools or combinations of tools and this knowledge is not being captured and codified and consequently not being exploited to its fullest.It is possible to build problem solving strategies around a tool or combinations of tools and codified knowledge about how to best utilise the tool/s.It should be possible to mix, match, combine, and exclude individual tools as components in a toolset employed by user defined problem solving strategies.Finally, it should be possible to automate the execution of user-defined strategies utilising tools and tool expertise.