Final Report Summary - MADCLADES (Cladogenesis and Niche Evolution in Madagascan Forests)
Nearly half of all plant species occur in tropical rain forest, a biome that only covers 7% of the world's terrestrial surface. In order to explain this peculiar concentration of plant diversity, we must understand eco-evolutionary dynamics (speciation, extinction, migration) happening on multimillion-year time scales. Data on the phylogenetic relationships of tropical rain forest plants are key to this understanding. To set the scene for the empirical part of MADCLADES, we searched the literature for hypotheses on the eco-evolutionary underpinnings tropical rain forest 'hyperdiversity' that have been, or can be, tested with phylogenetic data. Our review revealed a plethora of relatively ill-defined hypotheses, and showed that few studies have tested these hypotheses rigorously in a phylogenetic with phylogenetic data (Eiserhardt et al., in review). In this paper, we suggest a framework that integrates different hypotheses (Fig. 1), straighten out their definitions, and suggest how our knowledge of tropical rain forest hyperdiversity can be put on a better empirical basis. We expect this paper to be a key conceptual reference for future research on tropical rain forest diversity, and biodiversity patterns more generally. The empirical part of MADCLADES then addressed the shortage of phylogenetic studies of rain forest diversity by establishing a new model system for this purpose.
The largest impediment to understanding tropical rainforest hyperdiversity is the diversity itself, putting comprehensive studies out of reach. Instead, manageable yet representative model systems are needed. MADCLADES focused geographically on the island of Madagascar, which has been proposed as a "new model region" to study the formation of tropical biodiversity – yet plant studies are still few. Within Madagascar, MADCLADES focused on the palm family (Arecaceae), a globally established model group for tropical rain forest evolution that is diverse throughout Madagascar's humid forests. Being better served with biodiversity data than any other major plant group, Madagascan palms were by far the best option for creating a model for eco-evolutionary studies of the island's forests. However, prior to MADCLADES, a crucial element was missing: the evolutionary relationships (i.e. the phylogenetic tree) of Madagascan palms were essentially unknown. Inferring these relationships was a central aim of MADCLADES. Thanks to the project, a near-comprehensive (85% complete) phylogenetic tree of Madagascan palms will soon be available.
The ability of MADCLADES to deliver a phylogenetic tree of Madagascan palms depended critically on our ability to collect DNA samples in the wild. To this end, we secured complementary funding from multiple sources (National Geographic Society, International Palm Society, Royal Horticultural Society, Bentham-Moxon Trust, private donors), allowing us to conduct an extensive field campaign in Madagascar under the auspices of MADCLADES. To maximize the sampling of palm species for MADCLADES, we chose fieldwork locations based on species occurrence records and distribution models, targeting known species previously unsampled for DNA. Our Madagascan partners from the Kew Madagascar Conservation Centre, the University of Antananarivo, and the Zoological and Botanic Garden Tsimbazaza then assisted us with the legally and logistically difficult task of planning and conducting the field campaign.
Between November 2014 and December 2015, we conducted four field trips visiting 11 remote and palm-rich areas distributed across the lowland forest ecoregion of Madagascar (Fig. 2). The first trip covered a wide area in south-eastern Madagascar, including the localities of Ranomafana, Ifanadiana, Vatovavy, Manombo, Midongy Atsimo, and Vondrozo. On this trip, we were able to collect DNA samples of 31 palm species and a wealth of associated herbarium specimens, photographs and data. On the second trip, we visited the Marojejy massif in north-eastern Madagascar, collecting DNA and associated material of 18 palm species on a gradient from lowland rainforest to mountain shrubland. The third trip was conducted by our Madagascan collaborators, visiting the Mananara National Park, yielding DNA and associated material of 19 species. Finally, in November 2015, we conducted a major expedition to the remote and inaccessible eastern side of the Masoala peninsula, visiting the equally palm-rich western side on the way back. On this trip, we collected DNA samples and associated material of 34 palm species, greatly increasing our knowledge of this palm diversity hotspot. The extensive MADCLADES field campaign allowed us to reach our goal of near-comprehensive DNA sampling of Madagascan palms, essential for establishing the group as a model for evolutionary research. DNA samples and associated material will be available to other researchers through the Host Institution (Royal Botanic Gardens, Kew), and duplicates of herbarium specimens are also available at partner herbaria (TAN, MO, P).
Besides the DNA sampling, our fieldwork had other important outputs for the scientific community and beyond. During our campaign, we were able to collect five species unknown to science, including three massive tree palm species in eastern Masoala. These species are now being described scientifically as a result of MADCLADES. As part of the MADCLADES outreach activities, accounts of all four trips and the palm floras encountered will also be published in the peer-reviewed journal PALMS, read by scientists and amateurs alike (Baker et al. accepted, Baker et al. in review, Eiserhardt et al. in prep a,b). We also developed a field guide to the palms of the Marojejy massif (http://fieldguides.fieldmuseum.org/guides/guide/709) focusing on the summit trail, which is frequently used by tourists. On all trips, we worked closely with Madagascar National Parks guides, educating them in the local palm flora and its biological significance, enabling them to pass on this knowledge to other visitors. The joint fieldwork also allowed the Fellow to establish a firm collaboration with Madagascan colleagues at several institutions, which he intends to develop into further European-Madagascan projects.
To reconstruct the evolutionary relationships of Madagascan palms from DNA samples, MADCLADES opted for a cutting-edge phylogenomic approach, i.e. a combination of RNA hybrid baiting and genome skimming using the illumina sequencing platform. Phylogenomics using hybrid baiting was pioneered by MADCLADES collaborators (the Leebens-Mack lab at the University of Georgia, USA) using the American palm genus Sabal as a model. In 2014, the Fellow (W.L. Eiserhardt) visited the Leebens-Mack lab to be trained in this methodology and process a first set of samples. Subsequent lab work was carried out in the Jodrell laboratory at the Host Institution. As the hybrid baiting approach is spreading quickly and improvements are constantly being made, knowledge transfer is key to getting the best possible results. To this end, we founded the Palm Phylogeny Working Group at the World Palm Symposium in Colombia, 2015, through which we are now regularly exchanging experiences with several other labs. We also embedded the phylogenetic component of MADCLADES into the extensive Plant and Fungal Trees of Life research programme at the Host Institution, which uses a similar sequencing approach and engages actively in methodological developments. These links have not only benefited MADCLADES, but also allowed experiences from our project to impact the development of palm phylogenomics, and plant phylogenomics more broadly. The phylogenomic tree of Madagascan palms resulting from MADCLADES will be available shortly.
While working towards a robust phylogenomic tree of Madagascan palms, we contributed to the generation of a total-evidence phylogeny of the entire palm family (Faurby et al. 2016). Including all molecular evidence available from public repositories, plus information on morphology and taxonomy, this integrative "palm tree of life" represents the state of our knowledge of the evolutionary relationships of palms. The dual purpose of this tree is to serve as a provisional resource for eco-evolutionary palm research, while highlighting uncertainties and knowledge gaps, guiding targeted in-depth efforts such as MADCLADES.
The varied activities of MADCLADES allowed the Fellow to acquire a wide range of new skills covering four major areas of personal development: 1) working independently as a research leader, 2) collections-based research, 3) taxonomy, and 4) phylogenomics. The Fellow designed and conducted a major field campaign in a remote tropical area, providing him with critical leadership experience. To this end, he was coached by experienced Host Institution staff in e.g. fund raising, logistics, legal aspects (permits, risk assessments etc.), and interaction with local counterparts. Spending time in the field with experienced field workers allowed him to greatly improve his general fieldwork skills as well as specific expertise in collecting plant specimens (Fig. 3). This hands-on experience was complemented with formal training where needed (e.g. basic and advanced remote emergency care courses). To capitalize on the results of the field campaign, the Fellow was trained in palm taxonomy by the world experts, and acquired cutting-edge phylogenomic skills through hands-on training in two laboratories (at the Host Institution and at the University of Georgia, USA). Overall, the project provided the Fellow with an in-depth understanding of collections-based research at a world-leading institution, thus equipping him with the necessary experience for a career in this field.
During the course of the project, the Fellow represented MADCLADES with posters and talks at various international scientific meetings (UK Plant Evolution meeting 2014, Radiations conference Zürich 2014, European Network of Palm Scientists meeting 2013 and 2016, World Palm Symposium Colombia 2015), and further such presentations are planned. He was also involved in other outreach and engagement activities, such as teaching on the Plant and Fungal Taxonomy Diversity and Conservation MSc programme of the Host Institution and Queen Mary University of London. Partly because of his success with MADCLADES, the Fellow was able to secure a permanent position as Research Leader at the Host Institution. This will allow him to further disseminate the results of MADCLADES and fully exploit the potential of the data generated beyond the project's lifespan.