Extraction of Genic Interactions with the Recursive Logical Theory of an Ontology

Abstract

Abstract. We introduce an Information Extraction (IE) system which uses the logical theory of an ontology as a generalisation of the typical information extraction patterns to extract biological interactions from text. This provides inferences capabilities beyond current approaches: first, our system is able to handle multiple relations; second, it allows to handle dependencies between relations, by deriving new relations from the previously extracted ones, and using inference at a semantic level; third, it addresses recursive or mutually recursive rules. In this context, automatically acquiring the resources of an IE system becomes an ontology learning task: terms, synonyms, conceptual hierarchy, relational hierarchy, and the logical theory of the ontology have to be acquired. We focus on the last point, as learning the logical theory of an ontology, and a fortiori of a recursive one, remains a seldom studied problem. We validate our approach by using a relational learning algorithm, which handles recursion, to learn a recursive logical theory from a text corpus on the bacterium Bacillus subtilis. This theory achieves a good recall and precision for the ten defined semantic relations, reaching a global recall of 67.7% and a precision of 75.5%, but more importantly, it captures complex mutually recursive interactions which were implicitly encoded in the ontology.

BioNLP Shared Task - The Bacteria Track

“…The semantic annotation scheme was developed by two annotators through a series of independent annotations of the corpus, followed by reconciliation steps, which could involve concerted modifications [ 6 ]. As a third and final stage, the corpus was reviewed and the annotation was simplified to make it more appropriate to the contest.…”

“…Typically in the bacteria literature, the genic regulations are more likely expressed by the direct binding of the protein, while in the eukaryote literature, non-genic agents related to environmental conditions are much more frequent. The bacteria Gene Interaction task (GI) is based on [ 6 ], which is a semantic re-annotation of the LLL challenge corpus [ 1 ], for which the description of the GI events in a fine-grained representation includes the distinction between expression, transcription and other action events, as well as different transcription controls (e.g., regulon membership, promoter binding). The entities not only are protein agents and gene targets but also extend to families, complexes and DNA sites (binding sites, promoters) to better capture the complexity of the regulation at a molecular level.…”

Meta-Interpretive Learning of Data Transformation Programs

“…Similarly, ATRE has been used in [10] to learn a recursive logical theory of the ontology from a biological text corpus. However these ILP approaches cannot be easily extended for the general task of learning data transformation programs from examples.…”

“…However these ILP approaches cannot be easily extended for the general task of learning data transformation programs from examples. Both [2,10] have only been demonstrated on extraction of specific types of information from biological text, and did not attempt more general text transformation tasks of the kinds demonstrated in our paper. Moreover, these approaches were not shown to learn recursive rules from a small number of training examples and also do not support predicate invention.…”