-Xenbase (www.xenbase.org) is a knowledge base for researchers and biomedical scientists that employ Xenopus (X. laevis and X. tropicalis) as a model organism in gaining a deeper understanding of developmental and disease processes. Through expert curation and automated data provisioning from various sources this MOD (model organism database) strives to integrate the Xenopus body of knowledge together with the visualization of biologically significant interactions. We present a vision for the usage of various ontologies that facilitate the visualization of gene expression from a combination of experimental data sources and the linking thereof back to human disease modeling.
If you use Xenbase resources in your research please consider citing us, for example: Nucleic Acids Res. 2018 46(D1):D861-D868.
Major funding for Xenbase is provided by the National Institute of Child Health and Human Development, grant P41 HD064556.
In the development of the XPO a mapping of EQ
statements for Xenopus phenotype to the MP was used to
identify common statements to base the first version of the
XPO on. In keeping with the design patterns and
interoperability defined for phenotype ontologies, by applying
the structure of anatomy and physiology ontologies (Gkoutos
et al. [
Prior to the development of a phenotype curation interface in Xenbase. A Phenote version customized for the use of the XAO provided the curators with a means to explore phenotype annotations as applied to Xenopus related articles and images. This early work provided a wealth of insight into, and working with, pre/post composed statements. As an alternative to manual curation, especially for EaP (Expression as a Phenotype) the GEO (Gene Expression Omnibus) was identified as a data source for Xenopus RNA-seq and ChIP-seq raw data. A GEO metadata interface was developed to enable the curators to work with the GSE/GSM metadata to review the experiment and control samples together with the capturing of their background, manipulation and assay information. Once again utilizing ontologies including BFO, XAO and a subset of the PubChem small molecules identified in the Xenopus small molecules ontology (XSMO) as applicable to the experiments being curated. Through the NCBI eUtilitize the MINiML xml for the GEO metadata was sourced for Xenopus related experiments (approx. 220). The SRA run tables provided the link between the GSE, GSM and the raw data files. The bioinformatics pipeline for processing the RNA-seq or ChIPseq data was constructed in order to map to the current xenopus genome and to generate TPM read counts and DE (Differential Expression) reads based on the GEO metadata that had been reviewed and annotated by the curators.
The GEO loader is a java application developed to consume the GEO metadata as well as provide the staging of the TPM read counts and DE data. The population of a Star schema table structure in Xenbase is slated for the storage of this data in a Fact and Dimension table structure for subsequent consumption by the gene expression visualization that is being contemplated. The EaP statements derived from the DE reads would be stored in tables linked to the GSE and in turn the article and relevant image for the experiment.
The Insitu gene expression data that exists in Xenbase
provided the basis upon which to develop a modified heat map
to identify developmental stage and tissue expression of a gene
on a gene page. The TPM read counts (Session et al. [
Bridging the pre/post composition phenotype statements represents one of the issues that remains to be fleshed out. The application of a graph path to map one to the other and determine the accuracy, remains to be explored. The handling of copyright images (embargo) related to phenotype statements and the timing of the curation effort is another process flow topic that is under reviewed.
As the development of the XPO progresses and additional disease (DO) relationships are constructed, the benefits derived from the combined availability of expression as a phenotype, anatomical phenotype and gene ontology phenotype to contribute towards human disease modeling is anticipated. Additional data integration with Monarch via their dipper pipeline is also contemplated.