class: left, middle, inverse, title-slide # Biogeography of Polynesian Pteridophytes in a Global Context ### Joel Nitta, Alex White, Warren Wagner, & Eric Schuettpelz
National Museum of Natural History, Smithsonian Institution
### 3rd Annual Digital Data Conference 🌎
2019.06.10
https://joelnitta.com
--- ## Ever since Darwin... .pull-left-narrow[ Islands have been used as "natural experiments" to study evolution. - Few species - Many replicates - Clear hypotheses ] .pull-right-wide[ ![](https://natgeoeducationblog.files.wordpress.com/2015/05/galapagosfinches.jpg) .tiny[natgeoeducationblog.files.wordpress.com] ] --- ## Theory of Island Biogeography .pull-left-narrow[ Species richness as a factor of - island size - island isolation ] .pull-right[ ![](https://media1.shmoop.com/images/biology/biobook_biogeography_graphik_17.png) .tiny[media1.shmoop.com] ] .footnote[MacArthur & Wilson (1963, 1967)] --- ## Progression rule .pull-left[ Phylogeny recapitulates geology ] .pull-right[ ![](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4961166/bin/pnas.1601078113fig01.jpg) .tiny[Shaw & Gillespie (2016) *Proc Natl Acad Sci USA* 113] ] .footnote[Wagner & Funk (1995)] --- ## However... Most island studies focus on small clades within a single archipelago -- ... and that archipelago is usually Hawaii -- Few studies investigating patterns at **broader scales**, in other archipelagos --- ## Pteridophytes as a study system .pull-left-narrow[ Pteridophytes<br>(ferns and lycophytes)... - are over-represented on islands - play important ecological roles - need more study ] .pull-right-wide[ ![](https://www.fernsoftheworld.com/wp-content/uploads/2018/11/DSC_0205.jpg) .tiny[fernsoftheworld.com] ] --- ## Pteridophytes as a study system .pull-left-narrow[ Pteridophytes<br>(ferns and lycophytes)... - are over-represented on islands - play important ecological roles - need more study ] .pull-right-wide[ ![](https://www.fernsoftheworld.com/wp-content/uploads/2017/04/Dennstaedtia-Sundue-3183-4.jpg) .tiny[fernsoftheworld.com] ] --- ## Pteridophytes as a study system .pull-left-narrow[ Pteridophytes<br>(ferns and lycophytes)... - are over-represented on islands - play important ecological roles - need more study ] .pull-right-wide[ ![](https://www.fernsoftheworld.com/wp-content/uploads/2015/10/DSC_0172.jpg) .tiny[fernsoftheworld.com] ] --- ## Pteridophytes as a study system .pull-left-narrow[ Pteridophytes<br>(ferns and lycophytes)... - are over-represented on islands - play important ecological roles - need more study ] .pull-right-wide[ ![](https://www.fernsoftheworld.com/wp-content/uploads/2012/12/Asplenium-sp-Sundue-3003-4.jpg) .tiny[fernsoftheworld.com] ] --- ## Pteridophytes of Polynesia ![](index_files/figure-html/pacific-map-1.svg)<!-- --> --- ## Pteridophytes of Polynesia .pull-left[ - ca. 413 native spp. total - 141/182 <br> endemic/native to HI - 99/251 <br> endemic/native to FP - only ca. 20 native spp. in common ] .pull-right[ ![](https://www.dropbox.com/s/99hvplz1fhgad4l/tohiea.jpg?raw=1) ] --- ## Goals Trace dispersal **into** Polynesia Model diversification **within** Polynesia --- ## Goals Trace dispersal **into** Polynesia **⬅︎ Global scale** Model diversification **within** Polynesia **⬅ Regional scale** -- Need to assemble a dataset that can be used for **both** --- ## Data sources DNA sequence data: GenBank Occurrence data: - GBIF - Floras - Collections --- ## Data sources DNA sequence data: **GenBank** Occurrence data: - **GBIF** - **Floras** - **Collections** **Four** sources of names that need to be harmonized --- ## Taxonomic name resolution strategy Use Catalog of Life as taxonomic standard - single taxonomic concept - 13,994 accepted taxa - **43,599 synonyms** -- GenBank and GBIF: exact match on genus + species -- Floras and collections: fuzzy match on full scientific name -- Drop any records with names that can't be unambiguously resolved --- ## Occurrence data cleaning - All occurrences of pteridophytes (ferns and lycophytes) on GBIF:<br>9,422,314 initial records. - Use only records with GPS points, identified to extant species:<br>6,552,924 records kept. - Remove unusual records with CoordinateCleaner:<br>6,427,135 records kept. - Remove records with names that can't be resolved:<br>6,370,661 records kept. --- ## Phylogenetic analysis - Download *rbcL* for all pteridophytes on GenBank\*:<br>11,343 sequences / 5,024 species. - Resolve names, keep single best sequence per species:<br>4,150 species. - Infer phylogeny using maximum likelihood .footnote[\* `gbfetch` R package, https://github.com/joelnitta/gbfetch ] --- ## Biogeographic analysis - Infer **historical** biogeographic movements: DEC model in `BioGeoBears` (Matzke 2013) - Infer **extant** biogeographic structure: GoM model in `Ecostructure` (White et al. *Nature Comm.*, in press) --- class: middle, center # Results --- ## GBIF has good taxonomic representation of pteridophtyes <table> <thead> <tr> <th style="text-align:left;"> Source </th> <th style="text-align:right;"> Species </th> <th style="text-align:right;"> Genera </th> <th style="text-align:right;"> Families </th> </tr> </thead> <tbody> <tr> <td style="text-align:left;"> GBIF data </td> <td style="text-align:right;"> 10789 </td> <td style="text-align:right;"> 338 </td> <td style="text-align:right;"> 51 </td> </tr> <tr> <td style="text-align:left;"> Catalog of Life </td> <td style="text-align:right;"> 12996 </td> <td style="text-align:right;"> 340 </td> <td style="text-align:right;"> 51 </td> </tr> </tbody> </table> --- class: middle, center ## How do vouchered records compare to observations? 🤔 --- ## There are many more observation records, representing fewer species <img src="index_files/figure-html/compare-gbif-sources-1.svg" width="120%" /> --- ## Observation records are taxonomically skewed <img src="index_files/figure-html/gbif-records-hist-1.svg" width="120%" /> --- .tiny-margins[ ![](index_files/figure-html/abundance-map-1.svg)<!-- --> ] --- .tiny-margins[ ![](index_files/figure-html/richness-map-1.svg)<!-- --> ] --- background-image:url(radial_tree_with_realms_edited.png) background-size:contain background-position:50% 50% ## Mapping realms onto tree reveals NW/OW clades .footnote[ *rbcL* genus-level tree. Colors weighted by relative number of species per realm in each genus. Realms after Dinerstein et al. (2017) *BioSci*: 67. ] --- ## Classification of biological regions faces two challenges - Either classification must be made *a priori* - Or, clustering algorithms only allow for **discrete** membership --- ## A new method overcomes these challenges `ecostructure` (White et al. *Nature Comm.*, in press) - **estimates** clusters from the data - allows for membership in clusters to be **continuous** -- `STRUCTURE` (Pritchard et al. 2000) uses a similar framework to assign individuals continuously to genetic groups .center[<img src="https://www.researchgate.net/profile/Ozlem_Bulbul/publication/281140675/figure/fig1/AS:284586148941831@1444862033685/STRUCTURE-cluster-plots-for-the-K2-to-K5-range-of-population-structure-models-using-87.png" alt="drawing" height="200"/>] <br> .tiny[Bulbl et al. 2015] --- ## `ecostructure` ![](ecostructure_example_1.png) --- ## `ecostructure` ![](ecostructure_example_2.png) --- ## `ecostructure` ![](ecostructure_example_3.png) --- background-image:url(ecostructure_plot.png) background-size:contain background-position:50% 75% Pteridophytes of the world, *k* = 7 --- background-image:url(ecostructure_plot_highlight.png) background-size:contain background-position:50% 75% Pteridophytes of the world, *k* = 7 --- ## Pacific pteridophytes have affinities with SE Asia ![](index_files/figure-html/ecostructure-plot-3-1.svg)<!-- --> --- .center[ # Conclusions] -- - **Single** source of names with **comprehensive synonymy** is key to resolving taxonomy -- - Occurrence data are common in GBIF but should be used with caution -- - New world / old world splits a common theme in pteridophyte evolution -- - Pacific pteridophytes have affinities with SE Asia --- background-image:url(thank_you.png) background-size:cover background-position:50% 50% ## Thank you! Peter Buck Fellowship Smithsonian Institution DNA Barcode Network Pacific Tropical Botanical Garden <br><br><br><br><br><br> .white[ Jean-Yves Meyer (Délégation à la recherche, FP), Tom Ranker (UH), Ken Wood (NTBG), David Lorence (NTBG), Tim Flynn (NTBG), Greg Plunkett (NYBG), Mike Balick (NYBG), Ann Kitalong (Belau National Museum) ] --- class: middle, center # Extra Slides --- ## iNaturalist records not as common as I thought <img src="index_files/figure-html/check-gbif-obs-data-sources-1.svg" width="120%" />