Research (RIKEN) develops and maintains the RIKEN BioResource MetaDatabase. This database integrates several life science datasets to support researchers in making a comprehensive use of RIKEN's research results. For instance, RIKEN BRC collects, preserves and distributes various bioresources for further scientific research: experimental mouse strains, cultured cell lines and genetic materials of human and animal origin. In this article, we present a case study that mainly combines the RIKEN BioResource MetaDataBase [1] and the Bgee, a multi-species gene expression database [2]. Our use case involves finding the Alzheimer disease (AD) related human genes that are highly expressed in the healthy prefrontal cortex, and the RIKEN's genetically modified mice related to these genes. Information on the gene expression is obtained from the Bgee database and the human-mouse orthologs from the Orthologous MAtrix (OMA) database [3]. Orthologs are pairs of genes which have evolved from a single gene in the last common ancestor. The DisGeNET dataset provides relationships between disease and human genes [4].
Sciences. ∗Co-first authors. †Co-last authors.
APP
approaches for our use case. Although the query execution performance is significantly better in the centralised setup, unfortunately, it does not retrieve the latest results since it currently stores an older version of the Bgee database when compared to the data available via the remote Bgee SPARQL endpoint at https://bgee.org/sparql/. Moreover, the federated setup retrieves 12 more results than the centralised one. To avoid longer runtimes and query timeout, in both scenarios, we used the locally stored OMA and DisGeNET as named graphs in the RIKEN MetaDataBase.
Acknowledgments. Funding from State Secretariat for Education, Research and Innovation (SERI) via ETHZ grant BG 02-072020 and EU Horizon 2020 INODE grant 863410.