Attribute Exploration with Proper Premises and Incomplete Knowledge Applied to the Free Radical Theory of Ageing

摘要

The classical free radical theory of ageing assumes that oxidative damage by reactive oxygen species (ROS) accumulates with age in a self-enhancing process. The theory has been confirmed by many experiments in various species. However, it is seriously challenged since several years. In this ambiguous situation, we collected and ordered existing knowledge, with a focus on the integration of conflicting findings. We developed a specific method of knowledge base construction and give a first example of its application. Data reported in literature or generated by our experimental partners is formalized as Ripple Down Rules (RDR), a structure of general rules and exceptions. This rule set is validated and completed by the attribute exploration algorithm: Several, most specific RDR are accepted as background implications for an exploration starting from the examples collected during the RDR knowledge base growth. The RDR classify biological cases, which are defined by attributes like organism, cell type or stimulation experiment. The classes are different and chosen according to leading questions. We focus on low/high ROS concentration in age and on lifespan. Implications with proper premises are suited for such disjoint basic sets of premises and conclusions. We implemented an easily understandable exploration algorithm in conexp-clj, furthermore an extension of this algorithm to incomplete counterexamples. The correctness and completeness of both algorithms is proven.