Taxa

1
Taxa
IODE/JCOMM Forum on Oceanographic Data Management
and Exchange Standards

• Roy Lowry
• British Oceanographic Data Centre

2
Overview

• Definitions
• Use Cases
• Issues
• Taxon Lists
• WOD05
• ITIS
• WoRMS
• Recommendation

3
Definitions

• Taxon
• A taxonomic group of any rank (from species up to
  Kingdom)
• Taxonomic entity
• A type specimen or genome
• Taxonomy
• The hierarchical tree of taxa pertaining to a
  taxonomic entity

4
Use Cases

• Data value labelling (usage metadata population)
• Vocabulary requirements
• Stable tags
• Comprehensive coverage
• Access to change documentation (tag history)

5
Use Cases

• Semantic data aggregation
• Data marked up at species level
• Output required at higher level (e.g. diatoms,
  dinoflagellates etc.)
• Data discovery
• Smart discovery (ask for Copepoda, find Calanus
  finmarchicus
• Vocabulary requirements
• Comprehensive coverage
• Accessible taxonomy that may be integrated into
  local systems

6
Issues

• Biologists
• Biologists, especially taxonomists, have a very
  different mindset to information technologists
• Practices taxonomists view as progress are viewed
  in a very different light by the information
  technologist
• Renaming
• Reclassification
• Entity redefinition
• Taxonomic progress breaks simple data models
• Therefore we need something better

7
Issues

• Scale and dynamics
• There is a very, very large number of different
  entities that live in the sea
• Comprehensive cataloguing is a daunting task
• Every possibility of new entities being
  discovered
• Evolution is an ongoing process, sometimes at
  surprisingly rapid rates

8
Issues

• Uncertainty
• Taxon identification, particularly of preserved
  specimens, is an inexact science
• Not easy to decide whether the fuzzy green blob
  under the microscope is the same as the fuzzy
  green blob in the reference work
• Concept of identification confidence assumes
  importance
• Can manifest itself in taxon labels (e.g.
  affinis, confer)

9
Issues

• Identification Entity
• In an ideal world all taxa would be identified by
  linking to a known type specimen or genome
  (species level)
• Often identification only done to a higher level
  such as genus or even family
• Sometimes taxa are grouped together for counting
• Sometimes identification are morphological, not
  taxonomic (e.g. flagellate)

10
Issues

• Names or Codes?
• Database builders want entities labelled
  PERMANENTLY
• Biologists want to use names to label entities
• Biologists then want to change the names
• Result is much confusion and broken systems
• Database builders seem to be winning the day
  (even ERMS now has reference numbers)

11
WOD05 Taxon List

• Text file containing the following
• A significant subset of ITIS
• About 700 entities for taxa currently stuck in
  the ITIS ingestion process
• Codes for groups (e.g. nanoflagellates) that
  dont map to ITIS classification levels
• Codes for taxa combinations
• Codes for about 50 entities rejected by ITIS but
  used as labels in WOD
• Codes for a handful of biomass parameters

12
WOD05 Taxon List

• ITIS extension designed to address issues
• Incomplete cataloguing
• Imperfect identification entities
• Simple (flat file) technology fails to deliver
  access to entity relationships and classification
• Local extension of ITIS standard created to allow
  WOD to include all candidate data

13
ITIS

• Catalogue and fixed classification of taxa
• Some entries quality assured to a very high level
• Other entries (e.g. the entire USNODC taxonomic
  code list) used to establish initial population
  with no QC
• Very effective web query interface
• Full, up to date relational database (including
  taxonomy linkages) available for download

14
ITIS

• Good points
• Technology
• Quality of checked entries
• Bad points
• Coverage of marine organisms relatively poor
• Most marine organisms sourced from US data (e.g.
  about 10 of Baltic species present)
• Adding new entries takes for ever (2 years for a
  well-researched and presented application)
• Unwilling to take on global responsibility
• More interested in terrestrial than marine

15
WoRMS

• Catalogue and fixed classification of marine
  biological entities
• Many entries quality assured to a very high level
• Very effective web query interface
• Full, up to date relational database (including
  taxonomy linkages) available by request
• Ingestion from and mapping to ITIS (through
  entity codes yes Worms has them)

16
WoRMS

• Good points
• Technology
• Quality of checked entries
• Global ambition
• Consortium composition
• Marine focus
• Unknowns (by me)
• Coverage extent yet to be documented (but Im
  optimistic)
• Response to population extension requests as yet
  untested (again Im optimistic)
• Bad points
• No operational distribution mechanism for the
  full database

17
Recommendation

• WoRMS would seem to be the strongest candidate
  for a global marine taxa standard.