=Paper=
{{Paper
|id=Vol-1795/paper31
|storemode=property
|title=A Poster about a New Graph-based Taxonomy RESTful Data Service
|pdfUrl=https://ceur-ws.org/Vol-1795/paper31.pdf
|volume=Vol-1795
|authors=Leyla Jael García Castro,Leonardo Gonzales,Maria Martin
|dblpUrl=https://dblp.org/rec/conf/swat4ls/CastroGM16
}}
==A Poster about a New Graph-based Taxonomy RESTful Data Service==
https://ceur-ws.org/Vol-1795/paper31.pdf
A graph-based taxonomy RESTful data service
Leyla Garcia1*, Leonardo Gonzales1*, Maria Martin1
1 Protein Function Development Group, EMBL-EBI, Wellcome Genome Campus, CB10 1 SD,
Hinxton, UK
* These authors have equally contributed to this work
{ljgarcia, lgonzales, martin}@ebi.ac.uk
Abstract. In Life Sciences, taxonomy refers to a standard nomenclature and clas-
sification system for living organisms. In early days, organisms were clustered
together mainly based on shared morphological characteristics. Nowadays, evo-
lutionary patterns linking organisms have become more relevant when classify-
ing organisms. Applications and tools serving taxonomical data aim to facilitate
navigation. Here we present a graph-based RESTful data service built on top of
Neo4J. Our service supports common queries as well as advanced queries includ-
ing common ancestor between two or more taxonomic identifiers, a path to an-
cestors or descendants within a specified level, and the shortest path between two
taxa. Availability: http://www.ebi.ac.uk/proteins/api/doc/index.html
Keywords: Graph-based databases, taxonomy, web data services.
1 Background
Taxonomical data services aim to facilitate navigation through the hierarchy. They
commonly support queries based on taxon identifier as well as scientific and common
names. Responses include taxon identifier, scientific and common names, taxonomical
rank, e.g., genus, order or species, and lineage. For instance, the National Center for
Biotechnology Information (NCBI) provides taxonomy data services via e-utils tools,
supporting queries based on one or multiple taxonomical identifiers as well as scientific
and common names. Both Extended Mark-up Language (XML) and JavaScript Object
Notation (JSON) formats are supported as well as File Transfer Protocol (FTP) down-
loads and end-user interfaces.
The European Nucleotide Archive (ENA) provides a JSON-only taxonomy service,
supporting queries by taxon identifier and name. The name-based query can be user to
retrieve a list of suggested taxa based on the first characters. The Universal Protein
Resource (UniProt) serves taxonomical data as part of its Resource Description Frame-
work (RDF) distribution, accessible via its SPARQL Protocol and RDF Query Lan-
guage (SPARQL) endpoint. SPARQL flexibility together with SPARQL property paths
make it possible to create queries such as getting a path forwards or backwards from a
starting node. Both ENA and UniProt support FTP downloads and end-user interfaces.
Desirable features when traversing a taxonomical tree such as finding the shortest
path from one taxa to another or the common ancestor for multiple taxa are not com-
monly supported by traditional taxonomical data services. Here we present a graph-
based taxonomy RESTful data service supporting not only common queries but also
those related to taxonomical paths.
2 Our Neo4J-based RESTful taxonomy Service
Here we present a graph-based RESTful data service built on top of Neo4J. Our service
supports common queries –those related to taxonomic identifiers and scientific or com-
mon names, as well as advanced queries. In addition to the basic data, i.e., taxon iden-
tifier, scientific names, common names and lineage, we also provide links to the related
taxa. Advanced queries include the common ancestor between two or more taxonomic
identifiers, a path to ancestors or descendants within a specified level, and the shortest
path between two taxa. Such advanced queries might also be accomplished by using
SPARQL; however, they can turn out to be complex and long queries. In our service,
complexity is hidden to service users. Additionally, stress and load tests have been per-
formed on our service to ensure its reliability and robustness. Both XML and JSON
responses are supported. The full documentation is available at
http://www.ebi.ac.uk/proteins/api/doc/swagger/index.html?moduleId=taxonomyApi.
Some query samples are provided here:
Full data by taxon identifier, e.g., http://www.ebi.ac.uk/proteins/api/taxon-
omy/id/9606
Partial data by taxon identifier, i.e., basic information, children, parent or siblings,
e.g., http://www.ebi.ac.uk/proteins/api/taxonomy/id/9606/node
Lineage by taxon identifier, e.g., http://www.ebi.ac.uk/proteins/api/taxonomy/line-
age/9606
Full data by scientific, common and mnemonic names, e.g.,
http://www.ebi.ac.uk/proteins/api/taxonomy/name/homo?search-
Type=STARTSWITH&fieldName=SCIENTIFICNAME, useful for autocomplete
All taxa related with the queried taxon identifier in a specific direction and depth
level, e.g., http://www.ebi.ac.uk/proteins/api/taxon-
omy/path?id=8782&depth=3&direction=BOTTOM
Path between two taxa showing their relationship, e.g., http://www.ebi.ac.uk/pro-
teins/api/taxonomy/relationship?from=8782&to=9606
Lowest common ancestor of two or more taxa, e.g., http://www.ebi.ac.uk/pro-
teins/api/taxonomy/ancestor/8782,9606
Our graph-based RESTful taxonomy data service is a modern and robust service sup-
porting traditional queries as well as some new features that make it easier to find rela-
tionships across taxa. This service is part of the set of protein data web services deliv-
ered by the Protein Function Development Group at the European Molecular Biology
Laboratory – European Bioinformatics Institute (EMBL-EBI). We are exploring the
creation of a library of web components using and showcasing our web services.