Property Graph vs RDF Triple Store: A Comparison on Glycan Substructure Search

Abstract

Resource description framework (RDF) and Property Graph databases are emerging technologies that are used for storing graph-structured data. We compare these technologies through a molecular biology use case: glycan substructure search. Glycans are branched tree-like molecules composed of building blocks linked together by chemical bonds. The molecular structure of a glycan can be encoded into a direct acyclic graph where each node represents a building block and each edge serves as a chemical linkage between two building blocks. In this context, Graph databases are possible software solutions for storing glycan structures and Graph query languages, such as SPARQL and Cypher, can be used to perform a substructure search. Glycan substructure searching is an important feature for querying structure and experimental glycan databases and retrieving biologically meaningful data. This applies for example to identifying a region of the glycan recognised by a glycan binding protein (GBP). In this study, 19,404 glycan structures were selected from GlycomeDB (www.glycome-db.org) and modelled for being stored into a RDF triple store and a Property Graph. We then performed two different sets of searches and compared the query response times and the results from both technologies to assess performance and accuracy. The two implementations produced the same results, but interestingly we noted a difference in the query response times. Qualitative measures such as portability were also used to define further criteria for choosing the technology adapted to solving glycan substructure search and other comparable issues.

Comparing Relational and Ontological Triple Stores in Healthcare Domain

“…Various works also compare these RDF data stores [17][18][19][20][21]. In addition to comparing the RDF stores, RDF loaders are also compared in [22].…”

“…In addition to comparing the RDF stores, RDF loaders are also compared in [22]. Distinct from our work; [17][18][19] use generated data, [17,18] use insufficiently low data for the performance evaluation, [17][18][19][20]22] do not use medical data, and [19,22] compare both AllegroGraph and Oracle 12c. These studies use different metrics to query RDF data, but none of these studies compare the performance of triple stores using CONSTRUCT queries.…”

Knowledge Representation and Management: a Linked Data Perspective

“…Thus, literature will not be replaced by linked data but more data and knowledge can be easily expressed this way without hindering its accessibility. Besides RDF there are other graph-based models, such as Property Graphs [8], which may prove to be more effective in some specific areas of biomedicine.…”

Integration of Structured Biological Data Sources using Biological Expression Language

“…While RDF is highly expressive and each of these projects have developed and enforced well-defined schemata, the format is often not well-suited for downstream analyses and must first be queried with languages like SPARQL (SPARQL Query Language for RDF ) and subsequently be transformed into appropriate formats with 2 general-purpose programming languages. Alternatives to RDF/SPARQL such as property graphs (e.g., Neo4j , 3 OrientDB ) are comparable (Alocci et al , 2015) but also necessitate similar post-processing. 4 Conversely, there have been several biologically meaningful integration efforts (e.g., STRING; Warde-Farley, et al 2010, GeneMANIA; Szklarczyk et al , 2015, GeneCards;Stelzer et al , 2016 ).…”

“…Alternatives to RDF/SPARQL such as property graphs (e.g., Neo4j , 3 OrientDB ) are comparable (Alocci et al , 2015) but also necessitate similar post-processing. 4 Conversely, there have been several biologically meaningful integration efforts (e.g., STRING; Warde-Farley, et al 2010, GeneMANIA; Szklarczyk et al , 2015, GeneCards;Stelzer et al , 2016 ). However, most suffer from a lack of defined schemata or standardized data format that impede biological database interoperability.…”