[Taxacom] response to 60 new dragonfly species paper ("Rod thread")

Donat Agosti agosti at amnh.org
Fri Dec 18 09:05:00 CST 2015

Here a little technical update to KD's comment regarding the Odonatologica paper.

With the current workflows in place at Plazi (http://plazi.org) , here the status of the places where the 61 (!) species made it into cyberspace:

original PDF http://www.osmylus.com/images/own/Downloads/Odonatologica_44-4-low_res.pdf  
Biodiversity Literature Repository: DOI: 10.5281/zenodo.35388 http://dx.doi.org/10.5281/zenodo.35388   
Plazi TreatmentBank: Summary List http://plazi.cs.umb.edu/GgServer/summary/FF9B2A1CCA19FFEAEE22FFED4105FFB2  
Plazi treatments: RDF http://plazi.cs.umb.edu/GgServer/rdf/03A25264CA15FFFAEF37FB524211FE1F  
Plazi TreatmentBank: http://plazi.cs.umb.edu/GgServer/search 
Darwin Core Archive: DWCA http://plazi.cs.umb.edu/GgServer/dwca/FF9B2A1CCA19FFEAEE22FFED4105FFB2.zip  
Plazi Dashboards: https://drive.google.com/file/d/0B_yrQwn4yBySa3VBaGJwNDNXbnM/view?usp=sharing 
Zoobank entry http://zoobank.org/References/A0592344-0F17-4463-8CE2-02900DBB8F20  
Species-ID entry http://species-id.net/wiki/Umma_gumma  
Wikidata: https://www.wikidata.org/w/index.php?title=Special:WhatLinksHere/Q21714913&namespace=0&limit=100  
GBIF: record map http://www.gbif.org/occurrence/search?DATASET_KEY=7b04b312-ad6f-4161-b6a4-7d48bee99014  
NCBI Taxonomy Linkout: uploaded and processed, but not records (BOLD does not share these records)
MOL: in discussion
ORD.CH: in process (January)
EOL: in process (January)
EU-NOMEN: implementation in March 2016
BOLDSystems: request sent, no response

The entire workflow, or parts of exist, and will be part of the EU-BON project to import observation and trait data into the the planned European Biodiversity Observation Network. It is the first time that we processed a 240 page document with rich data included, and added all the new species as batch in Zoobank. All the treatments can now be cited, are available in different formats. The dashboards allow not only to visualize the content, but plays an invaluable tool for quality control of the data. The data can be queried on TreatmentBank.

This only could got to work because of a very productive (and enjoyable) collaboration with all the colleagues in the above mentioned institutions and their commitment to make it succeed.

The article got a DOI from the Biodiversity Literature Repository (http://biolitrepo.org ) at Zenodo. For one species (Umma gumma http://treatment.plazi.org/id/03A25264-CA15-FFFA-EF37-FB524211FE1F ) all the images have been uploaded to BLR as well and thus can be cited like the article itself.

Links to the barcodes provided in the (difficult to read) phylogenies to  BOLDSystems have been added.

The species have been linked to Wikidata and thus to a completely different audience.

Once the treatments are uploaded to Plazi, it takes a staggering, expected 15 minutes to get them published on GBIF!

The dashboards are another stunning way to visualize data in an article, immediately providing insights that are hardly obtainable otherwise, and also allow immediately to spot errors in the data, quantify the contributions by authors or institutions to such a revision. It also shows that number of specimens used in such a work, and from the distribution map it is possible to go back to individual treatments.

Limitations are that the BOLD and NCBI do not share their data (or only in a very limited sense), and thus it became obvious that the Barcode data is not in NCBI.
GBIF can accept the records, but since none of their taxonomy resources deal efficiently with new species, the species records are in GBIF, but not under the species name. As I understand, this is changing. This of course is relevant if GBIF wants to keep up with new species - generally the rare species - and those that have the most up to date identification, ie do not come along with the misidentifications marring some of the GBIF provided data.
The link from the specimen identifier to the digital object at RMNH could not be made.
As Puneet Kishor pointed out, this little project is like a romantic candle light dinner in a push place with many waiters and cooks taking care of the two guests. The issue is feeding hundreds of hungry mouth. One is to work on the ability to automate as much as possible in this workflow, which we are intended to do. BUT: the really important step will be to avoid creating this what cannot be described differently as mess. Not to publish anything that is beyond what technology has to offer, and where taxonomy is a leading domain in scientific publishing: Use open access and semantic enhanced publishing like it is implemented in the Biodiversity Data Journal http://biodiversitydatajournal.com/ , and to some degree in all the other Pensoft journals. The dashboards are the best illustration of the richness of the data that we miss out having a PDF only, and the integration of the article in the wider world (NCBI, GBIF) is really taking advantage of the Web. Finally, if we want to protect, the data has to be available, especially such rich and highly precise georeferenced data as provided in this article.

Looking at the content of this paper: This is an extremely rich, precise revision. Besides the detailed descriptions and illustrations, there are links to BOLD, and what is really stunning are the specimens included. Plotting them on a satellite image, their high degree of accuracies is just an amazing look, not to speak of an extreme importance to map life on earth. The authors statement is right: the real value is to spend the bulk part of the time in the field to collect new data

We really enjoyed working through this example and thank all the colleagues involved

Donat,  Jeremy and Guido


Dear taxonomic community,

Many thanks for the responses to our recent publication of 60 new African dragonfly species, which Ellinor kindly posted for me. I'm now able to react and give some background myself. Our primary purpose was to have the names available as quickly Dear taxonomic community,

Many thanks for the responses to our recent publication of 60 new African dragonfly species, which Ellinor kindly posted for me. I'm now able to react and give some background myself. Our primary purpose was to have the names available as quickly as possible, while once more pleading the case of the disappearing taxonomic expert. As comments focused on both the paper's form (let's call it the "Rod thread") and the media message ("Stephen thread"), and a lot of ground was covered, I'll respond to the latter in a later email.

On the personal side, my two co-authors have day jobs and study the insects they love in their free hours. I jump from one grant to the next odd job, as scientists often have to. The write-up was enabled by a six-month grant, half of which I reinvested in applying for new grants. The 60 names were needed for our ongoing efforts to get dragonflies applied more in freshwater conservation in Africa. Steps so far have been a species-level handbook, a complete Red List (probably the first for a tropical insect group), and all 127 000 records on GBIF (except for the 60 new species, as GBIF needs names first...). In the next months we'll get all 760 species'
identification, range maps, habitat details and indicator value online and give 80 trainees from 20 African countries their first introduction to freshwater entomology.

With three months to describe 60 species*, we opted for a stripped-down approach (please read the intro). GenBank and ZooBank are important platforms, but if I must chose between more taxonomic admin and getting those species assessed for the Red List or worked into a field guide, my priorities are clear. So with a "minimalist" approach, we wanted to demonstrate how to deal with the urgency. (Mind you, it's still 232 pages for 60 species!)

For example, why upload sequences to GenBank that are already on BOLD?
Also, if cyber-taxonomy can deliver the integration it promises, necessary data can be mined from a well-structured text. Computers can process data, but not discover species in the field or coin meaningful names. Indeed, the paper was immediately picked up by some helpful cyber-taxonomists to show how easily "traditional" rapid naming can be disclosed online: everything is now registered on Plazi and ZooBank, and ready for GBIF and Wikidata. We even had a small cyber-taxonomic revolution: for the first time unique identifier numbers generated in Plazi were accepted in Zoobank, rather than the reverse.

As said, the chosen approach was practical, but also intended to bare the challenges to taxonomic expertise today (more on that later). With dragonflies we are lucky that interest has actually boomed in the last decades, with a steady flow of new field guides. Interest from science and conservation is now increasing too, e.g. dragonflies are perhaps the most prominent insects in IUCN's red-listing efforts. However, this surge was realized largely by "amateurs". Our paper's support too came from this community. Not in the form of money, but interest. The choice to publish in Odonatologica (also run on private commitment!) was therefore obvious: give back to those who care most. So we ensured it would be accessible freely!

Cheers, KD

* Both paper and press release state that while 80% of Odonata species may be named, perhaps only a fraction of eukaryotes are. So while 60 species at once isn't exceptional in the grand scheme of things (e.g. beetles), it is for this specific order. 60 species add 1% to the known diversity in Odonata, the equivalent in Coleoptera is >3500. It adds 5% to what odonatologists think is left to name, which could equate to >25,000 in Coleoptera. So perhaps our job was indeed "easy"... ;o)

Klaas-Douwe 'KD' B. Dijkstra
Naturalis Biodiversity Center, Leiden, The Netherlands Conservation Ecology and Entomology, Stellenbosch University, South Africa science.naturalis.nl/dijkstra jrsbiodiversity.org/grant/stellenbosch_dragonflies/

Sixty new dragonfly species from Africa:
Press release: https://goo.gl/KGMsyC
Info and images: https://goo.gl/vRoJSL
Full publication: www.osmylus.com/index.php/downloads
Watch discovery of new species in DR Congo: youtu.be/Arr2k7dwzSU

Handbook of African Dragonflies:
Freshwater Biodiversity and Aquatic Insect Diversification:
Consensus classification of dragonflies:
Most complete damselfly phylogeny to date:
Taxacom Mailing List
Taxacom at mailman.nhm.ku.edu
The Taxacom Archive back to 1992 may be searched at: http://taxacom.markmail.org

Celebrating 28 years of Taxacom in 2015.

More information about the Taxacom mailing list