Computational protocol: The Use of Phylogeny to Interpret Cross-Cultural Patterns in Plant Use and Guide Medicinal Plant Discovery: An Example from Pterocarpus (Leguminosae)

Similar protocols

Protocol publication

[…] Total DNA was extracted from 0.2 to 0.3 g of leaf and/or flower tissue from herbarium or silica gel dried material using a modification of the Doyle and Doyle method . DNA was purified using QIAquick columns (Qiagen, Crawley, West Sussex, UK) following the manufacturer's protocol.The internal transcribed spacer 2 (ITS2), including parts of the 5.8S ribosomal RNA gene and the 26S ribosomal RNA gene, was amplified using primers ITS3 and ITS26E .The PCR protocol included a 2 min initial denaturation at 96°C and 32 cycles of 1 min denaturation (96°C), 1 min annealing (48°C), 50 s elongation (72°C), with a final elongation of 7 min at 72°C. The trnL-F intergenic spacer was amplified with primers “e” and “f” . The PCR protocol included a 4 min initial denaturation (96°C) and 32 cycles of 1 min denaturation (96°C), 1 min annealing (54°C), 1 min elongation (72°C) and final elongation of 7 min at 72°C. The barcoding fragment of matK was amplified with primers X and 3.2 . The PCR protocol included a 1 min initial denaturation (96°C) and 38 cycles of 30 s denaturation (96°C), 40 s annealing (46°C), 1 min elongation (72°C), with a final elongation of 7 min at 72°C. The first half of rbcL was amplified with primers rbcL1F and rbcL724R , following a protocol of 4 min initial denaturation (96°C), and 33 cycles of 1 min denaturation (96°C), 1 min annealing (50°C) and 1 min 20s elongation (72°C), with a final elongation of 7 min at 72°C. Finally, the ndhF-rpL32 intergenic spacer was amplified with primers ndhF and rpL32-R . Due to amplification of non-target product, we modified the PCR conditions given by as follows: one cycle of denaturation (96°C) for 2 min, 30 cycles of 95°C for 40 s, 52°C for 1 min and 65°C for 3 min 20 s with ramp of 0.3/s to 65°C and a final elongation cycle of 65°C for 5 min. All amplifications were performed in 30-µL volume reactions with BioMix (Bioline Ltd. London, UK).PCR purification and DNA sequencing of both strands were performed by Macrogen Inc. (Seoul, Korea). Complementary strands were assembled and edited with EditSeq (DNASTAR, Madison, WI). Alignments for rbcL and matK sequences were performed manually in BioEdit v. 7.0. ITS2, and the trnL-F and ndhF-rpL32 intergenic spacer sequences were aligned using CLUSTAL W , and adjustments were made manually in BioEdit v. 7.0, following the guidelines of Kelchner . All newly generated sequences have been submitted to GenBank (see and ) and the data matrix and phylogenetic tree generated here are available on TreeBase ( under the accession number 11586. [...] Sequence data were analysed under the Maximum Likelihood (ML) criterion, with RAxML using the partitioned model option with the GTR+Γ model and running 1000 bootstrap replicates .We borrowed two metrics from community ecology phylogenetics in order to assess and detect phylogenetic signal in medicinal properties. The first was the “comstruct” option in Phylocom 4.1 . This metric assesses the significance of phylogenetic signal for a community of taxa, which is the subset of a phylogeny. In other words, it calculates how significantly a group of species are clumped on the phylogeny. To do this, the mean phylogenetic distance (MPD) and mean nearest phylogenetic taxon distance (MNTD) for each sample (group of species on the phylogeny) is calculated and they are compared to MPD/MNTD values for randomly generated samples to provide p values for the significance of phylogenetic signal for the given sample (p values are calculated based on the frequency of random samples that were more clumped on the phylogeny than the real sample). For this study, we compiled “communities” of taxa that are used for one of the categories of use. This means that instead of grouping taxa based on which ecological zone or geographical area they are found, we grouped taxa that have similar uses in medicine together under one “community”. This way, we are able to assess the phylogenetic signal of each category of use on the phylogeny of Pterocarpus and answer the question: Are taxa used for a certain category more significantly related than expected by chance alone?The second metric used was the command “nodesig” in Phylocom v 4.1 . This option uses the same community sample as described above and tests each node of the phylogeny for overabundance of terminal taxa distal to it. Observed patterns for each sample are compared to those from random samples to provide significance for the observed overabundance. For a node that is identified through this approach, the descendants of this node are significantly more likely to belong to the “community” under consideration that expected by chance alone. As mentioned earlier, a “community” for this study represents the group of species used for a certain category of use. Hence, this technique identifies the exact position of phylogenetic clumping on the phylogeny, namely the “hot” nodes for a category of use. This can help us assess the predictive power of the phylogeny for the discovery of new medicinal species.The rationale behind using these metric is as follows: If a certain category of use shows strong phylogenetic signal, then closely related species demonstrate similar uses. With the first metric, we can asses which categories of use demonstrate strong phylogenetic signal. For these categories of use, we can subsequently identify which nodes on the phylogeny have more medicinal taxa than expected by chance, using the second tool. Taxa descending from these nodes are the ones that show significant “overabundance” in medicinal properties. Therefore, they deserve further investigation, including those species that are not reported in traditional medicine, as they are likely to share these properties with their relatives, as shown in . The matrix showing the samples used for all Phylocom analyses is given in .Analyses using these two approaches were carried out for each of the 13 categories of use mentioned above. Additionally, we performed the same analyses for three diseases of particular interest for which there is experimental evidence of bioactivity of Pterocarpus species: diabetes, malaria and cancer , , , , , , , , , , , . This also allowed a test of our methods at different levels of ethnomedicinal specificity (condition versus group of conditions). […]

Pipeline specifications

Software tools BioEdit, Clustal W, RAxML, Phylocom
Application Phylogenetics
Diseases Malaria