Making better species distribution models with improved spatial sampling and taxonomic knowledge

I am pleased to report that a paper I am a co-author on has just been published in Molecular Phylogenetics and Evolution. The work shows how species distribution models (SDMs), also referred to as ecological niche models (ENMs) can be improved both by using a phylogenetic framework to refine taxonomic units and also increasing spatial data (Figs. 1 and 2). As usual this is a collaboration between scientists across the world (Europe, Africa and the US), and is the second MPE paper from the PhD of my colleague Gabriela Bittencourt-Silva – congratulations!

The paper is entitled “Impact of species delimitation and sampling on niche models and phylogeographical inference: A case study of the East African reed frog Hyperolius substriatus Ahl, 1931″ and is available online here.


Screen Shot 2017-07-13 at 14.14.45.png
Fig. 1. Hyperolius substriatus samples used in this study (including samples from Lawson et al. 2013)
Screen Shot 2017-07-13 at 14.14.21.png
Fig. 2. ‘Splitting’ or ‘lumping’ the species distribution models can have dramatic effects on spatial predictions. Using greater numbers of sampling points with more accurate taxonomic information (from multi gene mitochondrial and nuclear DNA analyses) provide clearer insights into lineage distributions and phylogeography.

Vulnerable East African amphibian species makes the front page!

Screen Shot 2017-07-13 at 14.34.07.png

An image from work in our recent paper was selected as the cover of Diversity and Distributions – giving some much needed publicity to some of the beautiful amphibians and threatened forests that we research. This image is Afrixalus sylvaticus, listed as vulnerable on the IUCN red list and found in the Shimba hills of Kenya and forest patches of coastal Tanzania. Thanks to my colleague Beryl Akoth Bwong (now at National Museums of Kenya) for providing the nice photograph!

Evolution meeting 2017, Portland, Oregon

Portland logoThe Evolution meeting starts tomorrow! (as if you didn’t already know…).  Unfortunately I won’t be there in person, however my former PhD supervisor, Simon Loader (Natural History Museum London) will present a poster based on the paper I mentioned in the last post. I am sure Simon would be more than happy to talk about amphibian evolution and conservation in East Africa if you are interested. The poster is in the Comparative phylogeography poster session on Saturday 24th June.

Environmental correlates of phylogenetic endemism in amphibians and the conservation of refugia in the Coastal Forests of Eastern Africa

Our new paper is now available in early view in Diversity and Distributions here

With this work, my colleagues and I show how the inclusion of phylogenetic information for amphibians can help to identify priority areas for increased conservation efforts across Eastern Africa (Fig. 1).

Screen Shot 2017-06-01 at 12.22.17.png
Fig. 1. The Coastal Forests of Eastern Africa (CFEA), with the adjacent Eastern Afromontane (EA) biodiversity hotspot

Although species are arguably the most fundamental measure of biodiversity on earth, they often contain high levels of genetic and morphological diversity across their geographic ranges. Even with common organisms we may often see striking patterns of  differentiation below the species level which is often overlooked, and is especially relevant for local and regional conservation efforts. So how can we incorporate this intra-specific diversity into conservation assessment and planning? When we have sufficient data, we can build a molecular phylogeny (an evolutionary tree) to resolve the relationships of species, and include the major lineages that make up each of these species. Combining the geographic distribution of each of these lineages and species with its corresponding branch length on the phylogeny we can highlight areas where large amounts of evolutionary history have accumulated (see Fig. 2).


Screen Shot 2017-06-01 at 12.45.49.png
Fig. 2. A molecular phylogeny of species occurring in the coastal forests of Eastern Africa including species and intraspecific lineages, and a map of evolutionary history (PE) created by combining phylogenetic information with geographic data


This is the theoretical basis of our new paper, where we applied these methods for close to the full assemblage of coastal forest species across Tanzania and Kenya, leading to the compilation of the largest available genetic and geographic datasets to date. We hypothesised that many of the evolutionary hotspots in this region are important refugia for biodiversity in times of past climate change, and our results support the idea that long term climatic stability, benign current climate and topographic heterogeneity facilitate the persistence of evolutionary history (Fig. 4).

Screen Shot 2017-06-01 at 12.11.28.png
Fig. 4. Model contributions in the best GLM to explain PE

Finally, by intersecting our maps with the current protected area network we established that much of the region’s evolutionary history is poorly protected, and several areas should be prioritised for conservation in the future (Fig. 5).

Screen Shot 2017-06-01 at 12.11.37.png
Fig. 5. PE hotspots  intersected with the protected area network, showing the low levels of protection for much of the region’s amphibian evolutionary history



New accepted paper

Today a major manuscript that I have been working on for the past 3 years has been accepted in Diversity and Distributions – hurrah! Very happy to see this work published as it is of direct conservation relevance to the coastal forests of Eastern Africa. On top of that I submitted my PhD thesis two weeks ago – double hurrah!

I will post more updates on these two news items soon, until then I will be drinking a beer or two to celebrate…

Screen Shot 2017-07-13 at 14.43.20.png

International Biogeography Society biennial, Tucson, AZ

Last month I travelled to the US to attend the International Biogeography Society meeting in Tucson, Arizona where I gave a talk on my work investigating the biodiversity of the coastal forests of Eastern Africa using amphibians. The coastal forests are highly threatened and lacking rigorous assessment across them, especially involving genetic data. I used a method which identifies places that are special because they hold unique biodiversity and indicate refugia where biodiversity has persisted over time, while it has disappeared in surrounding areas (see the method here). I combined species distribution data with genetics to map geographic concentrations of evolutionary history and used environmental and historical climate data going back to the Last Glacial Maximum (~20,000 years ago) to see if the distribution of evolutionary history correlates with areas that have remained climatically stable over time. Results suggest that they do, and reinforces the idea that these parts of the coastal forest in East Africa are an important refuge for amphibians (and potentially other groups!) during times of severe climate change.

I find these kinds of methods really great because they enable us to pinpoint important areas to prioritise areas for conservation based on evolutionary history. This is particularly important given predicted future climate change and the impacts of deforestation in the tropics.


Bryce canyon, Utah


Sabino canyon, Arizona

New treefrog species from Ruvu South FR published

I posted a few months back that our paper about the new species from Ruvu South Forest Reserve near Dar es Salaam in Tanzania was in press. Now it’s published open access – you can download the paper for free here:

“A new, narrowly distributed, and critically endangered species of spiny-throated reed frog (Anura: Hyperoliidae) from a highly threatened coastal forest reserve in Tanzania”


Leapfrogging into new territory – the Mascarene ridged frog

Focusing on extraordinary organisms that are unique, or restricted to a particular area is always exciting, but sometimes even the seemingly ordinary can reveal fascinating insights into the historical biogeographic processes that have shaped the patterns of biodiversity we see today. Fairly common species found across a large area can tell us quite a lot about the effects of climate and evolution over time, especially those that have diversifed from a common ancestor into a complex of several related (but evolutionarily distinct) lineages or even species.

The Mascarene ridged frog Ptychadena mascareniensis is such an example, occuring across mainland and island areas of Africa mainly in savannah and open forest habitats and containing a number of divergent lineages with undescribed species diversity. Take a look at this new paper which investigates the evolutionary diversification of the P. mascareniensis species complex across its range and demonstrates how climatic niche evolution may have shaped current species diversity and distributions.

Graphical abstract copy.png

Fig. 1. Graphical abstract of the paper showing phylogenetic relationships, historical biogeography based on ancestral state reconstructions and bioclimatic niche of the P. mascareniensis complex.

Analyses within the paper show that there are at least ten distinct lineages of this species across Africa (7 in Africa, 3 on Madagascar, see Fig. 1) as a result of several speciation events mainly in the Miocene (over the past ~23 million years). Central Africa is identified as a diversity hotspot for these frogs with ‘out of Africa’ dispersal events to São Tomé in the west and Madagascar in the east. The P. mascareniensis from Tanzania where I focus my research appears to be the origin for the Madagascan radiation. The niches of these lineages are broadly similar in central Africa, with most tolerating similar climatic conditions (niche conservatism) but the lineages in West Africa and Madagascar exhibit very different ecological niches (niche divergence). The article is currently in press in Molecular Phylogenetics and Evolution.


A week of genomics and bioinformatics in Weggis

Last week I travelled to Weggis, Switzerland and attended the Bioinfomatics for Adaptation Genomics Winter School 2016 ran by the ETH Zurich. I was part of an international group of 30  postdocs and PhD students focused on learning some new  bioinformatic techniques to deal with NGS genomic data, which is a major part of my last PhD chapter. The five days of hands on bioinformatic analyses were absolutely great and introduced me to some software and concepts that I wasn’t yet experienced in – for example with full genome data (Erik Garrison), and GWAS (Arthur Korte).

However, the RADseq (Jon Puritz), detecting signatures of selection (Matteo Fumagalli) and inferring complex demographies (Daniel Wegmann) days were particularly relevant to my PhD project and an invaluable experience for me personally. I was extremely happy to hear the recommendations for designing NGS studies matched the major decisions that I made when I designed my own RADseq project, and looking around the room I could see that other people were also relieved to hear their experimental design was solid!

I have gained a huge amount of bioinformatic experience over the past year (mainly for RADseq data using Stacks and pyRAD) so it was great to hear about some other software I can use on my data (dDocent, ANGSD and NGStools, Fastsimcoal and ABCtoolbox). Overall I was extremely impressed with the skills and techniques we learned on this course; the organisers and instructors really did a great job, and it was massively beneficial. I would highly recommend this course to anyone wishing to take a peek inside the ‘black box’ of genomic analysis tools!