Learning Weighted Association Rules in Human Phenotype Ontology

Abstract

Human Phenotype Ontology (HPO) provides information about medically relevant phenotypes and the association of disease and phenotype concepts to HPO terms through annotations. The specificity of each HPO terms is estimated by its Information Content (IC), which assess the specificity of a term. An important research area focuses on the analysis of annotated data to extract knowledge. Association Rules (AR) can be used to discover relevant associations from annotated data. Classical AR methods consider all annotation equally, do not take into account that the HPO terms have different Information Content, i.e., different relevance. This implies the generation of association rules with low IC. This paper presents HPO-Miner (Human Phenotype Ontology-based weighted association rules), a methodology for extracting Weighted Association Rules from the HPO Ontology considering the IC of terms. To assess our methods, we tested HPO-Miner on publicly available HPO annotation datasets. The results demonstrate that our method outperforms the current state of the art approaches. HPO-Miner is publicly available at https://github.com/hguzzi/HpoMiner.