Rampant Reticulation in a Rapid Radiation of Tropical Trees - Insights from <i>Inga</i> (Fabaceae)

Schley, Rowan 1 ; Piñeiro, Rosalía2; Nicholls, James3; Pezzini, Flavia4; Farbos, Audrey1; Lewis, Gwilym P.5; Ringelberg, Jens6; Kidner, Catherine6; Twyford, Alex6; Dexter, Kyle6; Pennington, R. Toby1

Published Apr 08, 2025 on Dryad. https://doi.org/10.5061/dryad.tb2rbp09n

Data files

Apr 08, 2025 version files 2.19 GB

Data_1_MAFFT_Alignments.tar.gz
47.14 MB
Data_2_VCF_files.tar.gz
2.12 GB
Data_3_Codon_aware_Alignments.tar.gz
10.03 MB
Data_4_IQtrees.tar.gz
8.61 MB
README.md
2.89 KB

Abstract

Evolutionary radiations underlie much of the species diversity of life on Earth, particularly within the world’s most species-rich tree flora – that of the Amazon rainforest. Hybridisation occurs in many radiations, with effects ranging from homogenisation of species to generation of genetic and phenotypic novelty that fuels speciation, but the influence of hybridisation on Amazonian tree radiations has been little studied. We address this using the ubiquitous, species-rich, neotropical tree genus Inga, which typifies rapid radiations of rainforest trees. We assess patterns of gene tree incongruence to ascertain whether hybridisation was associated with rapid radiation in Inga. Given the importance of insect herbivory in structuring rainforest tree communities (and hence the potential for hybridisation to promote adaptation through admixture of defence traits), we also test whether introgression of loci underlying chemical defences against herbivory occurred during the radiation of Inga. Our phylogenomic analyses of 189/288 Inga species using >1300 target capture loci showed widespread introgression in Inga. Specifically, we found widespread phylogenetic incongruence explained by introgression, with phylogenetic networks recovering multiple introgression events across Inga and up to 20% of shared, likely introgressed, genetic variation between some species. In addition, most defence chemistry loci showed evidence of positive selection and marginally higher levels of introgression. Overall, our results suggest that introgression has occurred widely over the course of Inga’s history, likely facilitated by extensive dispersal across Amazonia, and that in some cases, introgression of chemical defence loci may influence adaptation in Inga.

https://doi.org/10.5061/dryad.tb2rbp09n

Description of the data and file structure

Data (Available on Dryad)

Data 1 - MAFFT alignments for all loci and all 3 datasets (Singlesp (one Inga accession per species only), Outgroup (A few Inga species, mostly Ingoid clade outgroups), and PPD (Singlesp dataset with paralogs removed)). Folder structure and file naming convention details dataset (e.g. Outgroup), hybseq locus (compxxxxxx) and alignment treatment (xxx_realigned).
Data 2 - VCF files used in Dsuite analyses for all 3 datasets (Singlesp, Outgroup, PPD). File naming convention details genus, method, dataset and filtering details (e.g. Inga_Dsuite_singlesp_DP8_Q20_filtered.vcf.gz)
Data 3 - Codon-aware alignments generated with OMM_MACSE for three subclade subsets within Inga, used for BUSTED analysis. Naming convention matches that of Data 1.
Data 4 - IQtrees generated for all locus alignments from all 3 datasets (Singlesp, Outgroup, PPD). Naming convention matches that of Data 1.

Supplementary Figures and Tables (Held as Supplemental Information on Zenodo)

Figures

Figure S1: Clade definitions for taxon subsets used in *f dM *and BUSTED analyses
Figure S2: Recovery success heatmaps
Figure S3ai: Inga single-accession-per-species ASTRAL tree
Figure S3aii: Inga single-accession-per-species concatenated IQtree
Figure S3b: Outgroup ASTRAL tree
Figure S3c: PPD ASTRAL tree
Figure S4a: Inga single-accession-per-species SplitsTree
Figure S4b: Outgroup SplitsTree
Figure S4c: PPD SplitsTree
Figure S5: PPD QC and DiscoVista plot
Figure S6a: Inga single-accession-per-species QD tree
Figure S6b: Outgroup QD tree
Figure S6c: PPD QD tree
Figure S7: Inga PPD D-statistic & F4 ratio plot
Figure S8a: Inga single-accession-per-species Fbranch
Figure S8b: Outgroup Fbranch
Figure S9ai: -loglikelihood h-value plot for single-accession-per-species PhyloNetworks run
Figure S9aii: -loglikelihood h-value plots for single-accession-per-species PhyloNetworks per-subclade run
Figure S9aiii: Inferred networks for single-accession-per-species per-subclade run
Figure S9b: -loglikelihood h-value plot for Outgroup PhyloNetworks run
Figure S10: Boxplots of f dM grouped by locus class and selection result for subclade subsets
Figure S11: Correlation plots of selection results grouped by locus class for subclade subsets

Tables

Table S1: Voucher information
Table S2: DiscoVista clade definitions
Table S3: Locus recovery success and alignment summaries
Table S4: -loglikelihood values per value of h for PhyloNetworks analyses
Table S5: BUSTED selection χ2 results
Table S6: ANCOVA BUSTED f dM results