This article is part of the supplement: Semantic Web Applications and Tools for Life Sciences (SWAT4LS) 2010
Argudas: lessons for argumentation in biology based on a gene expression use case
Citation and License
BMC Bioinformatics 2012, 13(Suppl 1):S8 doi:10.1186/1471-2105-13-S1-S8Published: 25 January 2012
In situ hybridisation gene expression information helps biologists identify where a gene is expressed. However, the databases that republish the experimental information online are often both incomplete and inconsistent. Non-monotonic reasoning can help resolve such difficulties - one such form of reasoning is computational argumentation. Essentially this involves asking a computer to debate (i.e. reason about) the validity of a particular statement. Arguments are produced for both sides - the statement is true and, the statement is false - then the most powerful argument is used. In this work the computer is asked to debate whether or not a gene is expressed in a particular mouse anatomical structure. The information generated during the debate can be passed to the biological end-user, enabling their own decision-making process.
This paper examines the evolution of a system, Argudas, which tests using computational argumentation in an in situ gene hybridisation gene expression use case. Argudas reasons using information extracted from several different online resources that publish gene expression information for the mouse. The development and evaluation of two prototypes is discussed. Throughout a number of issues shall be raised including the appropriateness of computational argumentation in biology and the challenges faced when integrating apparently similar online biological databases.
From the work described in this paper it is clear that for argumentation to be effective in the biological domain the argumentation community need to develop further the tools and resources they provide. Additionally, the biological community must tackle the incongruity between overlapping and adjacent resources, thus facilitating the integration and modelling of biological information. Finally, this work highlights both the importance of, and difficulty in creating, a good model of the domain.