https://doi.org/10.5061/dryad.47d7wm3f4

This README file provides information on the raw data, scripts, and results that accompany the paper “The symmetry spectrum in a hybridising, tropical group of rhododendrons” by Soza et al. 2022:

Soza, Valerie L., Ricardo Kriebel, Elizabeth Ramage, Benjamin D. Hall, and Alex D. Twyford. “The Symmetry Spectrum in a Hybridising, Tropical Group of Rhododendrons.” New Phytologist 243 (2022): 1491–1506. https://doi.org/10.1111/nph.18083.

The raw restriction-site associated DNA sequences (RADseq) that were used to generate consensus sequences representing loci within taxa are available at https://www.ncbi.nlm.nih.gov/sra/PRJNA758441.

Project name: Vireya RADseq

Executive Summary: We used restriction-site associated DNA sequencing (RADseq) on 114 taxa from Rhododendron sect. Schistanthe to reconstruct phylogenetic relationships and examine hybridisation. We then captured and quantified floral variation using geometric morphometric analyses, which we interpreted in a phylogenetic context.

Contact information regarding analyses: Valerie Soza, valeriesoza@gmail.com

NOTE: Version 1 published June 2022 was missing some image files. Version published August 2024 contains the full set of images (see folder “Outlines.zip”). No other changes were made between versions.

Methods of data processing and analysis:

Consensus sequences representing loci within taxa were assembled and aligned in ipyrad v.0.7.17. Four different datasets were generated that allowed different amounts of sample coverage across loci: requiring a minimum of 4, 37, 74 and 111 samples per locus. We refer to these datasets subsequently as minsamples4, minsamples37, minsamples74 and minsamples111 or min4, min37, min74 and min111, respectively, in filenames for alignments and phylogenetic trees deposited in Dryad. Resulting sequence alignments were analyzed by Maximum Likelihood in RAxML v.8.2.11 and the multispecies coalescent model in SVDQuartets for phylogenetic tree reconstructions. The RAxML min4 analysis was used to generate a chronogram in treePL. Introgression analyses were also performed using resulting alignments and trees with the D-statistic in Dsuite v.0.3r21. Use “Vireya_RADsamples_treenames_sed” file below for defining taxon codes in all alignment and tree files for full species names and accession details for Rhododendron sect. Schistanthe and outgroup samples used in analyses. See Soza et al 2022 for more detailed methods.

For morphometric analyses, we used images of the front and side views of flowers from sampled taxa above to quantify flower symmetry using elliptical Fourier analysis (eFa) in the R library Momocs. See Soza et al 2022 for more detailed methods.

Description of the data and file structure

The contents are as follows:

1.     R script

2.     SVDQuartets sequence alignments and phylogenetic analyses

3.     RAxML sequence alignments and phylogenetic analyses

4.     Chronogram and supporting phylogenetic information

5.     Dsuite results

6.     Morphometric analyses

7.     Vireya images

1. R script

“Rhododendron_morphometrics_dec2021.Rmd” is the R script used for the morphometric and shift detection analyses.

2. SVDQuartets sequence alignments and phylogenetic analyses

“reference-minsamples4woRlow3.u.snps.phy” is the sequence alignment of unlinked SNPs in phylip format for Rhododendron sect. Schistanthe sampled taxa that was generated from the min4 dataset, which required four samples per locus.

“reference-minsamples37woRlow3.u.snps.phy” is the sequence alignment of unlinked SNPs in phylip format for Rhododendron sect. Schistanthe sampled taxa that was generated from the min37 dataset, which required 37 samples per locus.

“reference-minsamples74woRlow3.u.snps.phy” is the sequence alignment of unlinked SNPs in phylip format for Rhododendron sect. Schistanthe sampled taxa that was generated from the min74 dataset, which required 74 samples per locus.

“reference-minsamples111woRlow3.u.snps.phy” is the sequence alignment of unlinked SNPs in phylip format for Rhododendron sect. Schistanthe sampled taxa that was generated from the min111 dataset, which required 111 samples per locus.

 “VirRAD1to6.R1.refmin4woRlow3.uSNPs.svd.allbs.con.rooted.newick2” is a rooted consensus phylogeny with bootstrap support for Rhododendron sect. Schistanthe generated using SVDquartets based on the alignment file “reference-minsamples4woRlow3.u.snps.phy”. This represents Figure 3 in Soza et al. 2022.

“VirRAD1to6.R1.refmin37woRlow3.uSNPs.svd.bs.con.rooted.newick2” is a rooted consensus phylogeny with bootstrap support for Rhododendron sect. Schistanthe generated using SVDquartets based on the alignment file “reference-minsamples37woRlow3.u.snps.phy”. This represents Figure S7 in Soza et al. 2022.

“VirRAD1to6.R1.refmin74woRlow3.uSNPs.svd.bs.con.rooted.newick2” is a rooted consensus phylogeny with bootstrap support for Rhododendron sect. Schistanthe generated using SVDquartets based on the alignment file “reference-minsamples74woRlow3.u.snps.phy”. This represents Figure S8 in Soza et al. 2022.

“VirRAD1to6.R1.refmin111woRlow3.uSNPs.svd.bs.con.rooted.newick2” is a rooted consensus phylogeny with bootstrap support for Rhododendron sect. Schistanthe generated using SVDquartets based on the alignment file “reference-minsamples111woRlow3.u.snps.phy”. This represents Figure S9 in Soza et al. 2022.

3. RAxML sequence alignments and phylogenetic analyses

“reference-minsamples4woRlow3.phy” is a PHYLIP format sequence alignment of concatenated loci for Rhododendron sect. Schistanthe sampled taxa and outgroups used in RAxML. This was generated from the min4 dataset, which required 4 samples per locus. This was used to generate the phylogeny in the file “RAxML_bipartitions.VireyaRADd10m5c91R1_0717_Rdref_minsamples4_fullmsa_rbsML2”.

“reference-minsamples37woRlow3.phy” is a PHYLIP format sequence alignment of concatenated loci for Rhododendron sect. Schistanthe sampled taxa and outgroups used in RAxML. This was generated from the min37 dataset, which required 37 samples per locus. This was used to generate the phylogeny in the file “RAxML_bipartitions.VireyaRADd10m5c91R1_0717_Rdref_minsamples37_fullmsa_rbsML”.

“reference-minsamples74woRlow3.phy” is a PHYLIP format sequence alignment of concatenated loci for Rhododendron sect. Schistanthe sampled taxa and outgroups used in RAxML. This was generated from the min74 dataset, which required 74 samples per locus. This was used to generate the phylogeny in the file “RAxML_bipartitions.VireyaRADd10m5c91R1_0717_Rdref_minsamples74_fullmsa_rbsML”.

“reference-minsamples111woRlow3.phy” is a PHYLIP format sequence alignment of concatenated loci for Rhododendron sect. Schistanthe sampled taxa and outgroups used in RAxML. This was generated from the min111 dataset, which required 111 samples per locus. This was used to generate the phylogeny in the file “RAxML_bipartitions.VireyaRADd10m5c91R1_0717_Rdref_minsamples111_fullmsa_rbsML”.

“RAxML_bipartitions.VireyaRADd10m5c91R1_0717_Rdref_minsamples4_fullmsa_rbsML2” is the maximum likelihood phylogeny from RAxML with bootstrap support for Rhododendron sect. Schistanthe generated from the min4 dataset, which required four samples per locus. This represents Figure S3 in Soza et al. 2022.

“RAxML_bipartitions.VireyaRADd10m5c91R1_0717_Rdref_minsamples37_fullmsa_rbsML” is the maximum likelihood phylogeny from RAxML with bootstrap support for Rhododendron sect. Schistanthe generated from the min37 dataset, which required 37 samples per locus. This represents Figure S4 in Soza et al. 2022.

“RAxML_bipartitions.VireyaRADd10m5c91R1_0717_Rdref_minsamples74_fullmsa_rbsML” is the maximum likelihood phylogeny from RAxML with bootstrap support for Rhododendron sect. Schistanthe generated from the min74 dataset, which required 74 samples per locus. This represents Figure S5 in Soza et al. 2022.

“RAxML_bipartitions.VireyaRADd10m5c91R1_0717_Rdref_minsamples111_fullmsa_rbsML” is the maximum likelihood phylogeny from RAxML with bootstrap support for Rhododendron sect. Schistanthe generated from the min111 dataset, which required 111 samples per locus. This represents Figure S6 in Soza et al. 2022.

4. Chronogram and supporting phylogenetic information

“VireyaRADd10m5c91R1_0717_Rdref_min4_raxml_treePLCIs.mean.newick” is the chronogram of Rhododendron sect. Schistanthe and outgroups displayed in Figure 5 in Soza et al. 2022 that shows branch lengths in million years. The chronogram was generated with treePL based on the tree topology from RAxML, using the min4 dataset.

“Vireya_RADsamples_tiplabels_ingroup_color_clade.xls” includes the geographic clades and flower colors for each species, used for data visualization.

“Vireya_RADsamples_treenames_sed” is an input file for the sed command to convert taxon codes in tree files to species names and accession details for Rhododendron sect. Schistanthe and outgroup samples used in phylogenomic analyses.

5. Dsuite analyses

“Rdreference.vcf” are all SNP data from the min4 dataset in vcf format that were used for estimating hybridization in Dsuite.

“VireyaRADd10m5c91R1_sets__D_BBAA_BHp” is the output of the D-statistic analysis for Rhododendron sect. *Schistanthe *using the BBAA option in Dsuite. This result is presented in Figure 4 in Soza et al. 2022.

 “VireyaRADd10m5c91R1_sets__Dmin_BHp” is the output of the D-statistic analysis for Rhododendron sect. *Schistanthe *using the Dmin option in Dsuite. This result is presented in Figure S10 in Soza et al. 2022.

“VireyaRADd10m5c91R1_sets__Dtree_BHp” is the output of the D-statistic analysis for Rhododendron sect. *Schistanthe *using the Dtree option in Dsuite. This result is presented in Figure S11 in Soza et al. 2022.

6. Morphometric analyses (Figures 6-7 and S12-S14 in Soza et al. 2022)

 “corolla_frontview_168.rds” is an R object file containing the morphometric data for the front view of flowers for all samples from Rhododendron sect. Schistanthe.

“corolla_sideview_166_landmarked.rds” is an R object file containing the morphometric data for the side view of flowers for all samples from Rhododendron sect. Schistanthe.

“Rhododendron_morphometrics.Rmd” is the R script to perform all morphometric analyses included in the manuscript.

“Rhododendron_morphometrics.pdf” is a PDF output of the morphometric analyses included in the R script “Rhododendron_morphometrics.Rmd”.

“corolla_frontview_168.rds” are the outlines of the front view of the corollas saved as a R .rds file.

“corolla_sideview_166_landmarked.rds” are the outlines of the side view of the corollas landmarked and saved as a R .rds file.

“shifts_front_PC1_to_PC3.rds” Results of shift detection analysis with data of the front view of the flowers.

“shifts_side_PC1_to_PC3.rds” Results of shift detection analysis with data of the side view of the flowers.

7. Vireya images (Figures 6-7 and S12-S14 in Soza et al. 2022) in folder “Outlines”.

334 image files of flowers used as raw data for the geometric morphometric analyses are provided with the following file name convention: species name (e.g. R. acrophilum), accession information (e.g. RBGE19801209 refers to a plant at the Royal Botanic Garden Edinburgh with accession number 19801209), photographer (e.g. ATwyford refers to Alex Twyford, VSoza refers to Valerie Soza), front/side indicates the image view.