Using Next Generation Sequencing to explore landscape-level processes in rainforest flora

Dr Hannah McPherson - Biodiversity Research Officer
 
Australian rainforests provide a unique model for the investigation of plant evolutionary patterns. The Bicentenary Plant Diversity Program is focused on combining genetic, ecological and environmental data to understand what the current distribution of species can tell us about how they responded to historical environmental and climatic changes. This will provide new insight on the evolution of the Australian flora as well as providing us with improved tools to predict the response of plants to current environmental and climatic modifications. Our research focuses on rainforest trees in order to explore patterns of differentiation in response to past climatic cycles at a landscape scale. Three sites at Washpool, Dorrigo, and Nightcap-Border Ranges provide the ideal landscape for our large scale comparative studies. Despite the sites being within 200 km of one another the assemblages of rainforest trees vary between the sites. Based on species assemblage data it is hypothesised that during past expansion and contraction cycles Nightcap provided a long-term refugium for many rainforest species, while Dorrigo and Washpool have lost species and been recolonised to varying degrees. Our study aims to address the question: What regulates the distribution and assemblage of northern NSW (subtropical) rainforest trees?

In total there are 370 rainforest tree species in Nightcap-Border Ranges, 253 species in Dorrigo and 146 in Washpool. Almost all the species at Dorrigo and all but one at Washpool are found in Nightcap and a total of 145 species occur at all three sites. Our sampling targeted approximately 100 rainforest tree species that are shared across the three sites in order to represent the landscape as broadly as possible.

Maurizio Rossetto, Rob Kooyman, Marlien van der Merwe and I collected approximately 2500 samples from the three areas for our molecular studies including population samples (up to ten individuals each) for each species from each site. In total we collected 126 species from Nightcap-Border Ranges, 104 from Washpool and 112 from Dorrigo. Approximately 10 individuals were collected from each of 70 species common to all three areas and about 90 species (10 individuals from each) occurred in at least two of three areas.

With the help of Katie Thurlby, Carolyn Porter, Nick Weston and Juelian Siow we have processed all 2500 silica-dried samples and DNA has been extracted by the Australian Genome Research Facility Ltd. In addition, voucher specimens from this program have greatly enhanced the rainforest collections in the National Herbarium of New South Wales. We have increased the number of rainforest species in the herbarium from Nightcap by 82%, Washpool by 91% and Dorrigo by 5%.

A molecular study that requires PCR optimisation and sequencing of multiple loci for multiple individuals for multiple species can be time consuming and expensive. One of the main aims of this program was therefore to explore and develop novel tools and approaches for addressing landscape-scale questions. With Next Generation Sequencing (NGS) becoming more accessible and cost-effective we conducted a pilot study using the Solexa platform from Illumina in order to explore the best NGS-based approaches for the broader study.

In this pilot study we have obtained shotgun sequence data from 13 Australian rainforest tree species, in eleven families, using whole genomic DNA. We constructed chloroplast genomes for each of these species and we are now investigating variation within and between the two distinct geographic regions. An example is Synoum glandulosum (family Meliaceae) for which we obtained shotgun data from two locations (North and South). For each location we pooled four individual samples. We trimmed the Illumina output for quality and length, mapped to a reference genome of Citrus sinensis (the closest relative according to APGIII that was available on Genbank) and assembled a chloroplast genome 155,119 base pairs (bp) in length. We identified variation between North and South samples corresponding to one single nucleotide polymorphism (SNP) in every 550 bp of chloroplast sequence. Although variation within each location (between the four pooled individuals) occurred at a much lower rate we found that the North harboured almost twice as much variation as the South.

The data have allowed us to explore the use of the chloroplast genome obtained through NGS for barcoding, SNP discovery and marker development. Preliminary investigations into the data reveal that nuclear sequence data are also present with sufficient coverage to be useful for assessing variation within and between populations and species of rainforest trees. We tested multiplexing, pooling and labelling of samples to determine the best approaches for the broader study. NGS is still in its infancy with regard to landscape-scale studies of wild populations so it will be critical to select the approach that maximises the number of species sequences and coverage of data while minimising cost. Certainly such approaches are the way of the future.

Stand of Nothofagus moorei. Photo: Maurizio Rossetto

Taking time to rest after a long day collecting: Hannah McPherson and Marlien van der Merwe. Photo: Maurizio Rossetto

NGS output showing variation within and between Synoum samples (top left), Synoum South mapped to Citrus sinensis reference genome (bottom left) and the Citrus sinensis chloroplast genome (right)