To obtain DNA sequence data for the chloroplast large subunit of ribulose 1, 5 bisphosphate carboxylase/oxygenase (rbcL) from New Zealand samples representing c. 200 genera for analyses in Biodiverse, and to use these data to undertake spatial analyses in Biodiverse.
There are 436 genera accepted in the New Zealand indigenous flora. Recent spatial and genetic analyses of the New Zealand flora used a generic level phylogeny to obtain genetic metrics for Biodiverse analyses (Heenan unpubl. data). Of the 436 genera included in the Biodiverse study, 214 genera were represented by sequences of species indigenous to New Zealand. However, 222 sequences were ‘surrogates’, being based on non-New Zealand species of the same genus or, in a small number of cases, a close generic relative.
This project was to obtain rbcL sequences from indigenous New Zealand species to replace sequences obtained from non-New Zealand samples and used as surrogates in the Biodiverse analyses. Spatial analyses utilising the new rbcL sequence data in the phylogenetic dataset will be undertaken.
4We obtained samples representing 211 indigenous New Zealand genera from dried collections in the Allan Herbarium or fresh collections from cultivated material. These 211 genera represented 95% of the 222 genera that were represented by surrogates in the Biodiverse study.
5Herbarium vouchers information is presented in Appendix 1.
DNA extraction and PCR
We extracted DNA from 309 samples and obtained suitable PCR product and clean sequences from 191 samples.
6For 118 samples we did not obtain sequences as we obtained no PCR product, weak PCR product, or the sequencing was messy.
7For some species we included more than one sample as we attempted multiple DNA extractions.
8Using a robot for DNA extractions has meant we were able to increase the sample number analysed, so we have been able to send three plates (each with c. 95 samples) for sequencing, rather than the two plates we initially envisaged. This has meant that for some samples that were unsuccessful in plate 1 (or plate 2) we were able to attempt another DNA extraction, PCR and sequence with a different sample in plates 2 and/or 3. For some genera we have attempted sequencing up to three different samples, and this has meant we have been able to obtain sequences representing more genera.
9For some samples identified as being particularly difficult, we attempted DNA extractions using a mortar & pestle and altered PCR protocols to obtain suitable DNA product for sequencing.
We obtained sequences for 191indigenous species that are representative of New Zealand indigenous genera; this is 86% of the target number of 222. For 167 samples we obtained read lengths of between 970 and the maximum of 1324 bases; for 10 samples there was a gap (32–223 bases) in the sequence between the internal primers; and for 14 samples we obtained only the 5′ or 3′ half of rbcL.
10Appendix 1 provides a summary of the successful sequence results, including GenBank numbers.
A data matrix was constructed comprising rbcL sequences representing 405 genera from this study and GenBank sequences that are based on indigenous New Zealand species. Sequences representing 31 genera that are based on non-New Zealand species acted as surrogates where there was not a sequence available based on a New Zealand indigenous species. One sequence was selected to represent each genus.
11The total dataset comprised 436 genera and was aligned in MEGA 5.0 (Tamura et al. 2011). A model of sequence evolution for rbcL was selected using ModelTest.
12An optimised Maximum Likelihood tree was used as the base tree for model likelihood calculations and the best model of sequence substitution was selected using the Bayesian information criterion. Bayesian inference of phylogeny was performed using MrBayes version 3.2.3 through the CIPRES Science Gateway version 3.3. Two runs with eight chains and a sample frequency of 5000 were run for 36,000,000 generations resulting in a total of 7200 trees for each run. The first 6000 trees of each run were discarded as burn-in and the remaining 2400 trees of both runs were combined in a 95% majority rule consensus tree using SumTrees version 3.3.1. The consensus tree is presented in Appendix 2.
13The Biodiverse software package version 1.0 was used for all analyses (Laffan et al., 2010). Spatial data used for this study comprised 213,141 georeferenced specimens from the New Zealand Virtual Herbarium (NZVH). All analyses were performed using a cell size of 0.12°, resulting in 2393 cells.
14The genus-level spatial data and phylogenetic tree were used to calculate phylogenetic diversity (PD) and phylogenetic corrected weighted endemism (PE_CWE) for the entire New Zealand archipelago and the main New Zealand islands. Statistical significance of the resulting patterns of endemism for each of the phylogenetic and non-phylogenetic analyses was assessed with a two-tailed test involving 999 random realisations of the observed datasets using the preserved model implemented in Biodiverse.
15Categorical Analyses of Neo- and Palaeo- Endemism (CANAPE) analyses. Phylogenetic diversity (PD) and phylogenetic weighted endemism (PWE) were calculated following Mishler et al. (2014) at genus and species rank. Statistical significance of the resulting biodiversity patterns for PD and PWE was assessed with 999 random realisations of the observed datasets using the preserved model implemented in Biodiverse. This model randomises the spatial locations of each taxon while preserving the taxon range and maintaining the taxon richness within each cell.
16A CANAPE analysis was performed on the genus and species spatial and phylogenetic data following Mishler et al. (2014). Differences in neo- and palaeo-endemism among islands were visualised as barplots by plotting the distribution of p(RPE) values for the entire New Zealand archipelago, North Island, South Island and offshore islands.