University of Glasgow

Roderic Page

We all have a \"past\" that we might not advertise widely, and my past
includes flirting with panbiogeography. Indeed my PhD thesis
[hdl:2292/1999](http://hdl.handle.net/2292/1999) is entitled
\"Panbiogeography: a cladistic approach.\" Shortly after graduating I
moved on to host-parasite cospeciation and the gene tree/species tree
problem (\"reconciled trees\", see Katz et al.
<http://dx.doi.org/10.1093/sysbio/sys026> for a recent example of this
approach), but part of me misses the glory days of vicariance,
dispersal, and panbiogeography.  
  
One thing which strikes me is how little use large-scale historical
biogeography makes of [GBIF](http://data.gbif.org) data. One of the
things that made Croizat\'s panbiogeography so interesting was the way
he exposed similar distribution patterns in unrelated groups of
organisms. He did this by hand, producing map after map, some
embellished with all manner of annotations (\"gates\", \"nodes\",
\"massings\", etc.). In some ways, Croizat as an early data miner. Now
we are awash in distributional data, where are the people revisiting
global scale historical patterns? In particular, wouldn\'t it be cool to
have a biogeographic search engine that could pull out taxa with
particular distribution patterns that we could then analyse.  
  
For example, while working on a project to map taxonomic names to
literature and genomics data, I embedded a widget to display GBIF maps.
Every so often I come across taxa have the classic \"Gondwana\"
distribution pattern. For example, below is a map for stoneflies of the
family the [Notonemouridae](http://data.gbif.org/species/7225/) from
GBIF.  
  
![](http://ogc.gbif.org/wms?request=GetMap&bgcolor=0x666698&styles=,,&layers=gbif:country_fill,gbif:tabDensityLayer,gbif:country_borders&srs=EPSG:4326&filter=()(%3CFilter%3E%3CPropertyIsEqualTo%3E%3CPropertyName%3Eurl%3C/PropertyName%3E%3CLiteral%3E%3C!%5BCDATA%5Bhttp%3A%2F%2Fdata.gbif.org%2Fmaplayer%2Ftaxon%2F7225%5D%5D%3E%3C/Literal%3E%3C/PropertyIsEqualTo%3E%3C/Filter%3E)()&width=721&height=362&Format=image/png&bbox=-180,-90,180,90){width="360"}  
  
Below is a map for the Notonemouridae using an orthographic projection
(see [earlier post for
details](http://iphylo.blogspot.com/2012/07/using-orthographic-projections-to-map.html)):  
  
![Notonemouridae](http://lh3.ggpht.com/-hYVjqmTNoZ8/UGA7m7GXkTI/AAAAAAAAB_g/CUJjJI0qD4c/Notonemouridae.png?imgmax=800 "Notonemouridae.png"){style="display:block; margin-left:auto; margin-right:auto;"
border="0" width="210" height="194"}  
Another family of stone flies, the
[Gripopterygidae](http://data.gbif.org/species/2999/), show a similar
pattern:  
  
![Gripopterygidae](http://lh4.ggpht.com/-5vUMpAmgc4k/UGA7mBXvBMI/AAAAAAAAB_Y/US8pgeljVTc/Gripopterygidae.png?imgmax=800 "Gripopterygidae.png"){style="display:block; margin-left:auto; margin-right:auto;"
border="0" width="201" height="193"}  
  
What I\'d like is to be able to query a database like GBIF for patterns
such as these Gondwanic distributions, then be able to pull out
associated phylogenetic information (e.g., via sequences in GenBank) so
that we could determine the antiquity of these patterns, and whether
they are consistent with geological models. We could begin to do
large-scale testing of biogeographic hypotheses in a (semi-)automated
way. At present we generally rely on a few well-studied examples that
are either broadly consistent with  

> Bocxlaer, I. V., Roelants, K., Biju, S. D., Nagaraju, J., & Bossuyt,
> F. (2006). Late Cretaceous Vicariance in Gondwanan Amphibians. (M.
> Hofreiter, Ed.)PLoS ONE, 1(1), e74.
> [doi:10.1371/journal.pone.0000074.t002](http://dx.doi.org/10.1371/journal.pone.0000074.t002)

  
or contradict  

> Cook, L. G., & Crisp, M. D. (2005). Not so ancient: the extant crown
> group of Nothofagus represents a post-Gondwanan radiation. Proceedings
> of the Royal Society B: Biological Sciences, 272(1580), 2535--2544.
> [doi:10.1098/rspb.2005.3219](http://dx.doi.org/10.1098/rspb.2005.3219)

  
the hypothesis that the history of biota of the southern hemisphere has
been largely structured by the break-up of Gondwana.  
  
A first step might be to index distributions at, say, family level and
above, and provide a series of polygons representing different
distribution patterns. We then search for distributions that are largely
concordant with those patterns, and query GenBank (or TreeBASE) for
sequences (or phylogenies) for those taxa. We then ask the questions
\"how old are these taxa?\" and \"what biogeographic histories do they
have?\"  
  


We all have a "past" that we might not advertise widely, and my past includes flirting with panbiogeography.

Messaggi di Rogue Scholar

Towards a biogeographic search engine