[Taxacom] progress on globalnames.org

[Peter DeVries]

I tried to send this earlier, but it seems to have bounced. Sorry if it is a
duplicate ...
---
I am working on some diagrams that explain this for a talk I am giving
Friday, so I will try to give you a somewhat simple version.

The first question is "Is there a need for a Species Concept Identifier"

If so, there are many different ways this could be implemented. Below is a
somewhat simplified version that should be handled as a separate issue,
since it would be too easy to dismiss question one by finding a flaw in the
implementation below. Also this particular species concept is somewhat
limited in scope just to be clear that it may or may not apply to all the
specimens of this particular species.

What do we mean when we talk about the specimens of Aedes triseriatus /
Ochlerotatus triseriatus collected in Wisconsin, Iowa and Minnesota?

We have a concept of a species, supported by studies indicating that this is
one metapopulation.

The species concept is somewhat tied to the type specimen in that the type
specimen and the type publication. They serve as starting points for the
species concept. However, until it is verified, the type specimen does serve
as a defining representation of this species concept.

The species concept is defined by a subset of specimens collected in
Wisconsin, Iowa and Michigan.

It is also defined by:

Photographs of those specimens
Occurrences of those specimens
Gene sequences from the paper mentioned previously
Publications involving those specimens

Now we have a somewhat robust model of what this species concept means.

Other workers have the choice of saying, I believe that these specimens
match the species concept known by this identifier (a Linked Data compliant
URL)

For instance, workers in northern Michigan believe that they have specimens
that are a match for the defined species concept. They then add that url tag
in the speciesconceptid field of their DarwinCoreTaxon record.

At some point in the future, those Michigan specimens are examined, and it
is decided that they should be added to the representative specimens that
define the species concept. A link is made in the species concept definition
back to the Michigan specimens.

The majority of specimens will link to the species concept, however, a much
smaller subset will serve as representative and have a link from the species
concept.

Now you have a much better foundation on which you can start to statements
and inferences about that species concept.

For instance, if Iowa researchers find that a particular gene is activated
when the mosquito is infected with West Nile Virus, you can infer with some
probability that that gene is also activated in that mosquito in northern
Michigan.

If a Wisconsin researcher discovers that that mosquito takes blood meals
from particular species of birds, you can infer that individuals in
Minnesota are also likely to take blood meals from the same bird species.

These statements are not always going to be accurate for all representatives
of that species concept, however, it is a lot more robust than what we
currently have.

Right now researchers often assume that if a statement was published about
Aedes triseriatus, that that statement applies to all specimens labeled as
Aedes triseriatus. However they miss any statements made about Ochlerotatus
triseriatus. Two different errors. Also, these statements are not machine
interpretable / analyzable.

We need species concepts to help better define what we are talking about and
allow large scale analysis. Species concepts are needed so we can tie
species data to all the earth science data that is becoming available in a
useful way. Otherwise we will not be able to ask the questions and solve the
problems that we need to.

Those species that are well understood and are the most important with
regards to data integration would be the first
to be available as web-accessible curated species concepts.

Respectfully,

Pete
------------------------------------------------------------
Pete DeVries
Department of Entomology
University of Wisconsin - Madison
445 Russell Laboratories
1630 Linden Drive
Madison, WI 53706
------------------------------------------------------------