Neuro-symbolic representation learning on biological knowledge graphs

Abstract

Motivation: Biological data and knowledge bases increasingly rely on Semantic Web technologies and the use of knowledge graphs for data integration, retrieval and federated queries. In the past years, feature learning methods that are applicable to graph-structured data are becoming available, but have not yet widely been applied and evaluated on structured biological knowledge. Results: We develop a novel method for feature learning on biological knowledge graphs. Our method combines symbolic methods, in particular knowledge representation using symbolic logic and automated reasoning, with neural networks to generate embeddings of nodes that encode for related information within knowledge graphs. Through the use of symbolic logic, these embeddings contain both explicit and implicit information. We apply these embeddings to the prediction of edges in the knowledge graph representing problems of function prediction, finding candidate genes of diseases, protein-protein interactions, or drug target relations, and demonstrate performance that matches and sometimes outperforms traditional approaches based on manually crafted features. Our method can be applied to any biological knowledge graph, and will thereby open up the increasing amount of Semantic Web based knowledge bases in biology to use in machine learning and data analytics.

Availability and implementation: https://github.com/bio-ontology-research-group/walking-rdf-and-owl.

Contact: robert.hoehndorf@kaust.edu.sa.

Supplementary information: Supplementary data are available at Bioinformatics online.

Fig. 1

Overview over the main steps in our workflow. We first build biological knowledge graphs by integrating Linked Data, biomedical ontologies and ontology-based annotations in a single, two-layered graph, then deductively close the graph using automated reasoning and apply feature learning on the inferred graph to take into account both explicitly represented data and inferred information. The two layers of the knowledge graph arise from the different semantics of linked biological data (represented in the graph-based language RDF) and the ontologies (represented in the model-theoretic language OWL); we formally connect the entities in the data layer through the rdf:type relation to ontology classes

Fig. 2

ROCAUC test scores of SIDER drug pairs the for predicting novel indications or targets or both