Enabling Web-scale data integration in biomedicine through Linked Open Data

Abstract

The biomedical data landscape is fragmented with several isolated, heterogeneous data and knowledge sources, which use varying formats, syntaxes, schemas, and entity notations, existing on the Web. Biomedical researchers face severe logistical and technical challenges to query, integrate, analyze, and visualize data from multiple diverse sources in the context of available biomedical knowledge. Semantic Web technologies and Linked Data principles may aid toward Web-scale semantic processing and data integration in biomedicine. The biomedical research community has been one of the earliest adopters of these technologies and principles to publish data and knowledge on the Web as linked graphs and ontologies, hence creating the Life Sciences Linked Open Data (LSLOD) cloud. In this paper, we provide our perspective on some opportunities proffered by the use of LSLOD to integrate biomedical data and knowledge in three domains: (1) pharmacology, (2) cancer research, and (3) infectious diseases. We will discuss some of the major challenges that hinder the wide-spread use and consumption of LSLOD by the biomedical research community. Finally, we provide a few technical solutions and insights that can address these challenges. Eventually, LSLOD can enable the development of scalable, intelligent infrastructures that support artificial intelligence methods for augmenting human intelligence to achieve better clinical outcomes for patients, to enhance the quality of biomedical research, and to improve our understanding of living systems.

An empirical meta-analysis of the life sciences linked open data on the web

“…Extracting schemas and vocabularies from the LSLOD cloud. For conducting our meta-analysis of the Life Sciences Linked Open Data (LSLOD) cloud, we have selected a set of LOD projects, SPARQL endpoints, and RDF graphs, by querying the metadata of the projects in the "Life Sciences" section of the popular Linked Open Data cloud diagram (April 2018 version) 23,30 , as well as through a literature review 13 of articles, documenting popular biomedical LOD projects, available on the PubMed search engine 4 . Furthermore, we established the following criteria for an LSLOD source to be included in the meta-analysis: (i) Each LSLOD source must have a functional SPARQL endpoint, (ii) For cases when an LSLOD source does not have a functional SPARQL endpoint, the source should be available as RDF data dumps that can be downloaded and stored in a local SPARQL repository, and (iii) Each LSLOD source must have at least 1,000 instances under any classification scheme that can be queried through the SPARQL endpoint.…”

“…In some cases, the same relation may be expressed in different RDF graphs using different semantics or graph patterns 13 . For example, the same attribute of molecular weight is expressed in two different LOD sources, DrugBank and KEGG, using drugbank:molecular-weight and kegg:mol_weight data properties respectively.…”

“…CHEMINF:000216 "average molecular weight descriptor" from the Chemical Information Ontology 47 where the • represents a blank node to capture additional information such as the source) RDF graphs. A user may wish to retrieve and reconcile such relations and attributes from both DrugBank and KEGG, as well as multiple other LOD graphs, to gain a complete picture in biomedicine, as unique drug-protein target relations may exist across the sources depending on the nature of their creation, or there may be conflicting information 13,40 . In the current state of the LSLOD cloud, the user must be aware of the underlying semantics and the data model used in these graphs as well as formulate lengthy SPARQL queries, which makes the entire process of information retrieval from the LSLOD cloud non-trivial.…”

“…Through a manual inspection of such "identifier" communities (e.g., communities 1417 and 1632 in www.nature.com/scientificdata www.nature.com/scientificdata/ sources. Most LOD querying architectures rely on the existence of exact URIs to query multiple sources simultaneously (using data warehousing, link traversal or query federation methods) 13,17,40 . However, since these URIs are essentially different, querying architectures will not be able to integrate data and knowledge from different sources (e.g., protein-ligand binding information from PDB and biological assay experimental data from PubChem).…”

“…Semantic Web and linked data technologies are deemed to be promising solutions to tackle these challenges, and enable toward Web-scale data and knowledge integration and semantic processing in several biomedical domains, such as pharmacology and cancer research [13][14][15][16][17] . Biomedical researchers have been some of the earliest adopters of these technologies, and have used them to drive knowledge management, semantic search, clinical decision support, rule-based decision making, enrichment analysis, data annotation, and data integration 12,16,[18][19][20] .…”