The effect of relational background knowledge on learning of protein three-dimensional fold signatures

As a form of Machine Learning the study of Inductive Logic Programming (ILP ) is motivated by a central belief: relational description languages are be tter tin terms of accuracy and understandability) than propositional ones f or certain real-world applications. This claim is investigated here for a p articular application in structural molecular biology, that of constructing readable descriptions of the major protein folds. To the authors' knowledg e Machine Learning has not previously been applied systematically to this t ask. In this application, the domain expert (third author) identified a nat ural divide between essentially propositional features and more structurall y-oriented relational ones. The following null hypotheses are tested: 1) fo r a given ILP system (Progol) provision of relational background knowledge does not increase predictive accuracy, 2) a good propositional learning sys tem (C5.0) without relational background knowledge will outperform Progol w ith relational background knowledge, 3) relational background knowledge doe s not produce improved explanatory insight. Null hypotheses 1) and 2) are b oth refuted on cross-validation results carried out over 20 of the most pop ulated protein folds. Hypothesis 3 is refuted by demonstration of various i nsightful rules discovered only in the relationally-oriented learned rules.