Main Page

From Protist-Prokaryote Symbiosis Database
Jump to navigation Jump to search


This project aims to describe symbiotic interactions between protists and prokaryotes as Linked Open Data. Information in the database is compiled from the scientific literature; over 500 symbiotic interactions described in 300 scientific publications are represented at present. The database is hosted on Wikibase Cloud, a hosting service for Wikibase instances provided by Wikimedia Deutschland.

What can the database be used for?

  • Search and browse symbiotic interactions by biological taxonomy, leveraging cross-references to external taxonomies and databases
  • Find interactions that were described in earlier literature but not yet studied with modern methods
  • Programatically find sequence data, literature, etc. by querying the NCBI databases using linked NCBI taxon IDs
  • Share data to be indexed in GloBI, through periodic data exports

What information do we wish to capture about symbiotic interactions?

Information Relevant database or ontology
Taxonomy of interacting organisms NCBI Taxonomy, Wikidata
Localization of symbionts in host organism Gene Ontology, Uberon
Nature of biotic interactions, if known/inferred Relation Ontology
Analytical methods used to identify organisms or interaction type OBI, Evidence Ontology
Environment where organisms were isolated Environment Ontology

Terms will be linked to other linked open data or ontologies, if there is a suitable exact match. The EMBL-EBI Ontology Lookup Service is a useful resource for browsing life science related ontologies.

This project originated as part of my doctoral dissertation (2017).

Similar projects elsewhere:

Documentation of the workflow and project administration is linked from Project:Main.

Explore the data

Some example entries to see how the data are modeled:

Each interaction statement is supported by one or more references to the scientific literature, linked by their DOI if available.

Use the Query Service (link on menu bar) to launch SPARQL queries; try the example queries to get started.

Q & A

Which types of interactions are in the scope of this database?
Interactions between protists (microbial eukaryotes) and prokaryotes (bacteria and archaea). In a few cases, interactions between two eukaryotic partners, or interactions with viruses, may also be included if they occur in organisms that also participate in protist-prokaryote interactions.
Fungi and macroalgae are generally not included, as are microbiome studies. Reports of putative symbionts described solely by morphology without further phylogenetic characterization may be included if the host organism is somehow notable, or if they are in older literature about host taxa that have not yet been studied with modern methods.
How are biological interactions between taxa represented in the database?
Taxa are represented as items in the database, linked together by statements (claims) representing the biological interaction. Each statement is supported by one or more references to the scientific literature.
Example: Pelomyxa palustris (Item:Q206) interacts with (Property:P19) Methanosaeta sp. ARCH (Item:Q203).
The item that goes before "interacts with" is the subject of the statement, while the item after it is the object.
In this database, we follow the convention that the subject of the "interacts with" statements is the "host" of the biological interaction, i.e. the physically larger organism in a pair, usually the eukaryote.
Further information about the interaction are encoded as "qualifier" statements. Attached to the example statement above is the following qualifier:
subject body part (Property:P20) cytoplasm (Item:Q29)
I.e. the symbiont Methanosaeta sp. ARCH is located in the cytoplasm of the host, Pelomyxa palustris.
How are taxa represented in the database?
Ideally, the taxa should be identified to the species/strain level, be associated with published sequence information, and be linked to external taxonomy databases. Sequence data is especially important for microorganisms, where morphology alone is insufficient to identify most species reliably. The associated sequence information provides an empirical reference point for disambiguation, should the taxonomy change in the future.
1. The taxon name/concept in the original scientific publication(s) correspond to the NCBI taxon, and sequence data are available.
Example: Candidatus Megaira polyxenophila is a bacterium that has been described in the scientific literature, and given a provisional ("Candidatus") name at the species level. An NCBI taxon ID has been assigned to this taxon name, which corresponds to the taxon concept from the literature.
In this database, such items are encoded as instances of the class "taxon".
In some cases, a taxon concept has been assigned a different name in the publication from the name that appears in the NCBI Taxonomy record. This usually happens when a sequence was first published in NCBI under a temporary name, which has not been updated with the subsequent formal name. These are nonetheless encoded here as "taxon" items, but are additionally tagged as "NCBI taxon name needs updating". Example: Candidatus Vesiculincola pelomyxae is still listed under a provisional name "Eubacteriales bacterium SKADARSKE-1" in the NCBI Taxonomy, but the provisional name is clearly unique and refers to the same taxon entity as the published name.
2. The taxon name/concept in the original publication does not correspond exactly to the NCBI taxon, but sequence data are available.
Example: The ciliate species Eufolliculina methanicola was described and named in a scientific publication, but sequences from that study were published under a non-specific identifier "Folluculinidae sp." in the NCBI Taxonomy, that potentially represents sequences from multiple taxa.
Therefore, the item for Eufolliculina methanicola is encoded in this database as an instance of "placeholder taxon", because it does not have an exact equivalent in the NCBI Taxonomy, although we think it represents a legitimate taxon, based on the information in the cited references.
To allow us to track the identity, should the NCBI Taxonomy be updated in the future, a representative NCBI sequence record for this taxon, as given in the original publication that described it, is linked to this item with the property "representative sequence for placeholder taxon".
3. No sequence data are available for a given taxon name/concept.
Possible reasons:
  • Sequence data were produced in a study but not published
  • Organism was identified by morphology alone, without using sequencing methods
Such items are encoded as instances of "named organism without published sequence data".
Example: A species of Arcobacter was identified as a symbiont of Bihospites bacati, but although sequencing of a marker gene was reported, the sequence was not published.
4. The organism is only identified to a higher taxonomic level.
Possible reasons:
  • Organism was identified by morphology alone, without using sequencing methods
  • Organism was identified with group-specific molecular probes, but these were not sufficient to identify them to the species or strain level.
Example: The ciliate Frontonia leucas is associated with unclassified Alphaproteobacteria, identified with group-specific molecular probes. However, no sequence data were reported that would allow the bacteria to be identified more specifically.
Such items are also encoded as instances of "named organism without published sequence data".
Item class/subclass in this database Sequence data available NCBI Taxon equivalent Wikidata item equivalent taxon rank
taxon yes yes maybe species or strain
placeholder taxon yes no maybe species or strain
named organsim without published sequence data no no maybe any
How is the taxonomic hierarchy represented in the database?
The taxon items linked by "interacts with" statements should be at the species or strain level, similar to the concept of a "submittable organism name" in the ENA database (
Where possible, taxon items are linked to their corresponding NCBI taxon ID and/or Wikidata item, so they can be placed into the context of either the NCBI Taxonomy or the Wikidata taxonomy trees.
However, not all taxa have corresponding NCBI taxon IDs or Wikidata items. In some cases ("placeholder taxon"), the NCBI Taxonomy may be inconsistent with the published literature or not yet updated. In other cases ("named organism without published sequence data"), there is information about their taxonomy, but sequence data are not available. Or the taxon may not meet the notability criteria for inclusion in the Wikidata database (e.g. strains or species that have not been formally named).
Therefore, every taxon item is linked to its next highest formally named parent taxon. The parent taxon item is linked to Wikidata. This allows us to query the database taxonomically using the Wikidata taxonomy tree, with a federated SPARQL query, even for taxon items that are not represented in external databases.
The parent taxon items are instances of the class "higher taxon" and should not be used for "interacts with" statements.
Is this a standalone project? Why not directly integrate this into Wikidata?
This project was built as a standalone Wikibase instance, instead of directly adding entries to Wikidata, because:
  • Not all taxa meet the notability criteria for Wikidata, e.g. species that have not been formally named
  • We wish to structure the concepts and statements different from Wikidata, e.g. using "interacts with" statements
  • Our preferred classification may conflict with what is already present in Wikidata
Nonetheless, this is not a standalone project as items are cross-referenced to Wikidata, NCBI Taxonomy, and other ontologies, to permit federated queries. Where possible, taxa that were previously not represented in Wikidata and which meet their notability criteria have been added there. We have also proactively informed the NCBI Taxonomy team about taxon names that required updating in their database.
Why use a single 'interacts with' statement, with qualifiers for interaction type, instead of different properties for each interaction type?
The nature of an interaction is often not fully understood, or may have multiple facets. Coding interaction types as qualifiers allows us to stack multiple functional roles on a single interaction
Why host this on Wikibase?
This database has seen a number of iterations: starting as a table in a word processor file, to spreadsheets, a custom SQLite database, and an attempt to homebrew a structured data base with XML files and Python scripts. After getting some experience on Wikidata, I found that Wikibase offers the key features that I wanted: flexible and extensible schemata, graphical frontend for manual data entry, options for programmatic data import from tables, integration with external databases, and a sophisticated query interface.
What is the beautiful organism depicted in the logo?
Kentrophoros sp. H
(The logo may not be visible in the mobile version of this site.)