Data and code from: Local adaptation to current but not future climate in seasonally tropical dry forests: Population genomic evidence in a Neotropical legume tree

Data files

Jul 27, 2026 version files 112.67 KB

R_script_future_climate_models.R

47.79 KB
README.md

64.88 KB

Abstract

The resources provided here comprise data sets used in our study "Local adaptation to current but not future climate in seasonally tropical dry forests: population genomic evidence in a Neotropical legume tree" comprising genomic data from double digest restriction-site associated DNA (ddRAD) sequencing (i.e. 29266 SNPs) for 109 individuals of Enterolobium cyclocarpum sampled across 12 sites in Colombia. We also provide climate raster layers for the species in northern South America from global climate models for 4 different climatic variables (i.e., temperature seasonality, precipitation of the driest month, average monthly precipitation during the warmest quarter of the year, and average monthly mean temperature over the driest quarter of the year) using three time periods (2041–2060, 2061–2080, and 2071–2100), two emission scenarios (SSP2-4.5 and SSP5-8.5) and three different global climate models ("EC-Earth3-Veg-LR", "GFDL-CM4", and "MPI-ESM1-2-LR"). The R scripts used to reproduce all figures and analyses presented in the paper, including population structure, detection of candidate SNPs, partial RDA, isolation by distance and environment, and genomic offsets, among others, are provided.

Dataset DOI: 10.5061/dryad.vhhmgqp8n

Description:

This dataset contains files to analyze genotype–environment associations, patterns of local adaptation, isolation by distance and by the environment, relative gene flow, and genomic offsets under future climate scenarios as described in the paper.

The R script "R_script_future_climate_models.R" is hosted on Dryad. The zipped files "Files_to_run_R_script.zip", "Qgis_Enterolobium.zip", "SNMF_analysis_files.zip", and "SNP_calling_and_filtering.zip" are hosted on Zenodo.

Parameters M, n, and m:

"M": Allowed number of mismatches between putative alleles to define putative loci

"n": The allowed number of matches between putative loci during catalog construction

"m": Minimum number of raw reads required to form a stack or putative allele

Files:

1) "SNP_calling_and_filtering.zip"

This compressed folder contains scripts used to run the Stacks pipeline and VCFtools for SNP genotyping using demultiplexed ddRAD sequences. Hosted by Zenodo.

File	Description
"denovo_test_M_n_and_m.sh" (9 files numbered from 1 to 9).	Bash scripts used to run the de novo Stacks pipeline as a test to detect the best combination of parameters, M, n and m from demultiplexed ddRAD-seq data (stored in the ENA nucleotide repository - accession number PRJEB112779). The script uses a subset of samples specified in the file "pop_test_b.txt". The script runs "denovo_map.pl" for M=n from 1 to 9, and m from 1 to 9 (i.e. 9 scripts). The scripts then execute the "population" module for loci present in at leat 80% of the samples (i.e. -r =0.8). The output includes different population parameters including the number of sites (i.e. SNPs), variant sites (i.e. variant SNPs), polymorphic sites (i.e. polymorphic SNPs), and % of polymorphic loci (i.e. % of polymorphic SNPs). After executing the scripts, polymorphic sites obtained with the different combinations of M=n from 1 to 9, and m = 1 to 9 are compared. The parameters M=n=m equal to 4 maximized the amount of polymorphic SNP loci. An Excel file with the summary of polymorphic SNP loci obtained with different combinations of M, n and m is provided.
"filtered_data_set_denovomap_M_n_m_4.sh"	Bash script used to run the de novo Stacks pipeline (denovo_map.pl) for SNP genotyping using demultiplexed ddRAD sequences corresponding to 109 samples, using M, n, and m equal to 4. This script uses the file "pop_test_filtered.txt" to assign populations (based on sampling sites) to each of the 109 samples. After executing the bash script a raw VCF file containing SNP genotypes is generated.
"VCFtools_Enterolobium.sh"	Bash script used to filter the raw VCF file using VCFtools with the following settings: removal of indel sites (i.e., – remove-indels), allowing sites with a maximum of 50% missing data (i.e., – max-missing 0.5), including only bi-allelic sites (i.e., – min-alleles 2 and – max-alleles 2), sites with mean read depths between 5 and 100 (i.e., – min-meanDP 5, and – max-meanDP 100), and genotypes with mean read depths between 5 and 100 (i.e., – minDP 5, and – maxDP 100).
"populations_module_Enterolobium.sh"	Bash script used to run the populations module in Stacks on the filtered VCF dataset to retain loci present across 20% of the samples, and generate output files in VCF and PLINK formats for downstream analyses.
"pop_test_a_filtered.txt"	Text file containing the IDs of 110 samples (109 of those used in downstream analyses) and their initial population assignments based on sampling sites. First column contain original sample names with IDs, and the second column (i.e. initial population assignment) indicates abbreviated sampling localities: antio , vallc , pnare, tolim, vicha, casan, dmeta,cesar, minge, minca, luriz,and charc, which correspond to localities: Loc. 10, Loc. 11, Loc. 7, Loc. 6, Loc. 12, Loc. 9, Loc. 8, Loc. 5, Loc. 4, Loc. 3, Loc. 2, Loc. 1., as detailed in Figure 2 and Table S2. This file is used by Stacks to specify which demultiplexed samples are included to build the SNP catalog using M = n = m = 4, the combination maximizing the amount of polymorphic SNPs.
"pop_test_b.txt"	Text file containing the IDs of a subset of 36 samples used during parameter optimization tests in the de novo stacks pipeline, assigning all samples to a single population. First column contain original sample names with IDs for 36 selected samples, and the second column (i.e. population assignment) indicates "singlepop" (i.e. all samples assigned to the same population). This subset was used to evaluate different combinations of M, n, and m parameters and identify the combination maximizing the number of polymorphic SNPs.
denovo_test_m3_m4_m5_m6_m7_m8.xlsx	Excel file with the summary of polymorphic SNP loci obtained with different combinations of M, n and m.

2) "R_script_future_climate_models.R"

R script used to generate future climate scenarios for for bio4 (temperature seasonality), bio9 (average monthly mean temperature over the driest quarter of the year), bio14 (precipitation of the driest month), and bio18 (average precipitation during the warmest quarter of the year), under different global climate models for the periods 2041–2060, 2061–2080, and 2071–2100, using two scenarios (i.e. "SSP245" and "SSP585"),and three different global climate models (i.e. "EC-Earth3-Veg-LR", "GFDL-CM4" and "MPI-ESM1-2-LR"). The use of a virtual machine or a computer with high RAM memory is recommended to run the content of the script. Hosted by Dryad.

3) “SNMF_analysis_files.zip”

This compressed .zip file contains the R script and associated input and output files used for population genomic structure analyses using "sparse non-negative matrix factorization" (SNMF) as implemented in the LEA package. Hosted by Zenodo.

File	Description
"SNMF_R_script.R"	This R script performs population structure analysis on RADseq genomic data from samples of Enterolobium Cyclocarpum by running SNMF (sparse Non-negative Matrix Factorization) across K = 1–10 clusters using the "LEA" package. The script also calculates cross-entropy scores to determine the best number of clusters, and extracts ancestry proportion matrices (Q-matrices) for K = 2, 3, and 4 to build stacked barplots using the package "ggplot2" for each K value (with individuals ordered by their most likely cluster assignment). At the end of the script the admixture barplots (i.e. from K=2 to 4) are combined and plotted into a single figure. The script also saves ancestry coefficients as CSV files to re-run the script and generate admixture barplots for K = 2 to 4 without the need to re-run the SNMF analysis.
"output_no_tolima_no_bad_apples.ped"	Gentoype file in ".ped" format containing genotypes for 109 samples after discarding low quality samples. Genotypes are provided for 29266 SNP loci and are coded as "0" (i.e. presence of two reference alleles), "1" (i.e. presence of one alternate and one reference allele), or "2" (i.e. presence of two alternate alleles). "9" denotes missing genotypes. This file is used to obtain a file of genotypes in ".geno" format, which is the format read by the "LEA" package.
"output_no_tolima_no_bad_apples.geno"	Gentoype file in ".geno" format containing genotypes for 109 samples obtained from the ".ped" file.
"SNMF_ancestry_K2_January_13_2026.csv"	Q-matrix (ancestry proportion matrix) for K =2 saved in ".csv" format. Variables V1 and V2 represent the estimated ancestry proportions for the Caribbean and Orinoquía genetic clusters, respectively. The "Sample Name" column corresponds to the code assigned to each sample, as indicated in Figure 2 and Table S1 of the manuscript.
"SNMF_ancestry_K3_January_13_2026.csv"	Q-matrix (ancestry proportion matrix) for K = 3 saved in ".csv" format.Variables V1, V2 and V3 represent the estimated ancestry proportions for the Caribbean, Orinoquía, and Cauca Valley genetic clusters, respectively. The "Sample Name" column corresponds to the code assigned to each sample, as indicated in Figure 2 and Table S1 of the manuscript.
"SNMF_ancestry_K4_January_13_2026.csv"	Q-matrix (ancestry proportion matrix) for K = 4 saved in ".csv" format. Variables V1, V2, V3 and V4 represent the estimated ancestry proportions for the Caribbean, Upper Magdalena Valley, Orinoquía, and and Cauca Valley genetic clusters, respectively. The "Sample Name" column corresponds to the code assigned to each sample, as indicated in Figure 2 and Table S1 of the manuscript.
"ordered_samples.csv"	This ".csv" file is implemented in the "SNMF_R_script.R" as a reference to arrange individuals in the same order across admixture barplots for K=2-4.

4) "Files_to_run_R_script.zip"

This compressed folder includes all files to run the core analyses. Included files are the R script to run the core analyses (i.e. "Script_Genomic_Analyses_Enterolobium.R"), SNP genotype datasets (i.e., PLINK “.raw” and VCF “.vcf” files) for 109 individuals, allele frequency datasets for specific SNP subsets, environmental predictors based on bioclimatic variables under future climate scenarios as described in the paper, and spatial layers used to define geographic regions to plot genomic offset results (i.e. Caribbean, inter-Andean valleys, Orinoquía region). A description of each file is provided below. Hosted by Zenodo.

Main R script file

File

Description

"Script_Genomic_Analyses_Enterolobium.R"

This R script contains the full workflow to run the core analyses. The pipeline is structured into the following steps which are commented and better described within the R script: 1) Setup. 2) Processing of genomic data. 3) Detection of SNPs associated with neutral alleles. 4) Processing of environmental data. 5) Principal Components used as proxies for population structure. 6) Partial redundancy analysis (RDA) controlling for population structure and geography. 7) Identification of candidate SNPs. 8) Visualization of RDA results. 9) Visualization of SNP loadings. 10) Classification of adaptive alleles at candidate SNPs based on the correlation obtained with each climatic predictor. 11) Calculation of per-individual frequencies of candidate SNPs carrying adaptive alleles associated with heat or drought stress. 12) Visualization of of scaled climate variables and scaled per-individual frequencies of SNPs carrying adaptive alleles associated with heat or drought stress. 13) Clines of candidate SNPs carrying adaptive alleles (per-individual) associated with heat or drought stress. 14) Population allele frequency change at candidate SNPs across climatic gradients using generalized additive models (GAMs). 15) Preparation of datasets for genomic offset analyses. 16) Estimation of genomic offsets. 17) Relative gene flow analyses using "divMigrate". 18) Analyses of isolation by distance and isolation by environment.

CHELSA climatic and sample information files

Files

Description

"Chelsea_Enterolobium_for_RAD_no_tolima_ind"

Long term (i.e. covering the period 1981-2010) climatological means taken from the CHELSA climatic data portal (https://www.chelsa-climate.org/datasets/chelsa_climatologies) for bio4 (temperature seasonality) - column TS, bio9 (average monthly mean temperature over the driest quarter of the year) - column TDQ, bio14 (precipitation of the driest month) - column PDM, and bio18 (average precipitation during the warmest quarter of the year) - column PWQ, for 109 individuals. The column named "Coordinates_for_RDA.ID_reduced" correspond to sample IDs described in Table S1. Columns named "Lat" and "Long" correspond to geographical latitude and longitude.

"pop_test_filtered.txt"

Text file containing regional and population assignments for each sampled individual.

Genomic data files

Files	Description
"plink_raw_file_no_tolima_no_bad_apples_r20.raw"	Genotype matrix in ".raw" PLINK format for 109 individuals and 29,271 SNPs. Genotypes are coded as 0, 1, or 2, representing the number of alternative alleles. Missing or excluded genotypes are coded as NA.
"r20_m5100_miss0.50_2alleles_no_bad _apples_no_tolima_ind_.recode.p.snps.vcf"	SNP genotype dataset in ."vcf" format (variant call format) including information on the reference and alternative alleles.

Allele frequency datasets used to estimate genomic offsets

Files	Description
"t_ent_frequencies_2026_all_candidates.csv"	Population allele frequencies at candidate SNPs (i.e. 1568 SNPs). The first column "chr_snp" corresponds to a given allele at a candidate SNP loci (those loci are detailed in the Supp. File 1). Columns "sp01antio", "sp01vallc", "sp01pnare", "sp01tolim", "sp01vicha", "sp01casan", "sp01dmeta", "sp01cesar", "sp01minge", "sp01minca", "sp01luriz", "sp01charc" correspond to each population at the twelve sampling localities, respectively: Loc. 10, Loc. 11, Loc. 7, Loc. 6, Loc. 12, Loc. 9, Loc. 8, Loc. 5, Loc. 4, Loc. 3, Loc. 2, Loc. 1., as detailed in Figure 2 and Table S2.
"t_ent_frequencies_2026_heatdrought_candidates.csv"	Population allele frequencies at candidate SNPs (i.e. 1568 SNPs). The first column "chr_snp" corresponds to a given allele at a candidate SNP loci (those loci are detailed in the Supp. File 1). Columns "sp01antio", "sp01vallc", "sp01pnare", "sp01tolim", "sp01vicha", "sp01casan", "sp01dmeta", "sp01cesar", "sp01minge", "sp01minca", "sp01luriz", "sp01charc" correspond to each population at the twelve sampling localities, respectively: Loc. 10, Loc. 11, Loc. 7, Loc. 6, Loc. 12, Loc. 9, Loc. 8, Loc. 5, Loc. 4, Loc. 3, Loc. 2, Loc. 1., as detailed in Figure 2 and Table S2.
"t_ent_frequencies_2026_non_adaptive_same_ lenght_as_adaptive_candidates_heat_drought_stress.csv"	Population allele frequencies at candidate SNPs (i.e. 1568 SNPs). The first column "chr_snp" corresponds to a given allele at a candidate SNP loci (those loci are detailed in the Supp. File 1). Columns "sp01antio", "sp01vallc", "sp01pnare", "sp01tolim", "sp01vicha", "sp01casan", "sp01dmeta", "sp01cesar", "sp01minge", "sp01minca", "sp01luriz", "sp01charc" correspond to each population at the twelve sampling localities, respectively: Loc. 10, Loc. 11, Loc. 7, Loc. 6, Loc. 12, Loc. 9, Loc. 8, Loc. 5, Loc. 4, Loc. 3, Loc. 2, Loc. 1., as detailed in Figure 2 and Table S2.
"t_ent_frequencies_2026_set_non_adaptive.csv"	Population allele frequencies at candidate SNPs (i.e. 1568 SNPs). The first column "chr_snp" corresponds to a given allele at a candidate SNP loci (those loci are detailed in the Supp. File 1). Columns "sp01antio", "sp01vallc", "sp01pnare", "sp01tolim", "sp01vicha", "sp01casan", "sp01dmeta", "sp01cesar", "sp01minge", "sp01minca", "sp01luriz", "sp01charc" correspond to each population at the twelve sampling localities, respectively: Loc. 10, Loc. 11, Loc. 7, Loc. 6, Loc. 12, Loc. 9, Loc. 8, Loc. 5, Loc. 4, Loc. 3, Loc. 2, Loc. 1., as detailed in Figure 2 and Table S2.

Climatic data under CMIP6 models

Future climate scenario raster layers (downscaled) representing bioclimatic variables derived from CMIP6 global climate models. Files are provided in paired ASCII grid format (.grd and .gri). These files are needed to run all genomic offset analyses. Each raster file follows the naming convention "bio_[Climatic variable][Time period][Emission scenario]_[Global Climate Model]_ag_10.grd/.gri". As ".grd" and ".gri" files are complementary they must be kept together in the same working directory. A description of the files is provided below.

Files	Description
"bio4_1981_2010_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio4_1981_2010_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio4_1981_2010_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO4 (i.e. temperature seasonality) in ".grd" and ".gri" format. Files correspond to the time-period 1981–2010, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio4_1981_2010_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio4_1981_2010_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio4_1981_2010_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO4 (i.e. temperature seasonality) in ".grd" and ".gri" format. Files correspond to the time-period 1981–2010, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio4_2041_2060_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio4_2041_2060_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio4_2041_2060_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO4 (i.e. temperature seasonality) in ".grd" and ".gri" format. Files correspond to the time-period 2041–2060, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio4_2041_2060_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio4_2041_2060_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio4_2041_2060_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO4 (i.e. temperature seasonality) in ".grd" and ".gri" format. Files correspond to the time-period 2041–2060, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio4_2061_2080_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio4_2061_2080_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio4_2061_2080_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO4 (i.e. temperature seasonality), in ".grd" and ".gri" format. Files correspond to the time-period 2061–2080, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio4_2061_2080_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio4_2061_2080_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio4_2061_2080_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO4 (i.e. temperature seasonality) in ".grd" and ".gri" format. Files correspond to the time-period 2061–2080, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio4_2071_2100_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio4_2071_2100_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio4_2071_2100_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO4 (i.e. temperature seasonality) in ".grd" and ".gri" format. Files correspond to the time-period 2071–2100, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio4_2071_2100_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio4_2071_2100_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio4_2071_2100_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO4 (i.e. temperature seasonality) in ".grd" and ".gri" format. Files correspond to the time-period 2071–2100, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio9_1981_2010_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio9_1981_2010_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio9_1981_2010_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO9 (i.e. average monthly mean temperature over the driest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 1981–2010, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio9_1981_2010_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio9_1981_2010_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio9_1981_2010_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO9 (i.e. average monthly mean temperature over the driest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 1981–2010, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio9_2041_2060_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio9_2041_2060_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio9_2041_2060_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO9 (i.e. average monthly mean temperature over the driest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 2041-2060, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio9_2041_2060_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio9_2041_2060_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio9_2041_2060_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO9 (i.e. average monthly mean temperature over the driest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 2041–2060, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio9_2061_2080_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio9_2061_2080_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio9_2061_2080_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO9 (i.e. average monthly mean temperature over the driest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 2061-2080, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio9_2061_2080_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio9_2061_2080_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio9_2061_2080_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO9 (i.e. average monthly mean temperature over the driest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 2061-2080, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio9_2071_2100_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio9_2071_2100_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio9_2071_2100_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO9 (i.e. average monthly mean temperature over the driest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 2071-2100, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio9_2071_2100_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio9_2071_2100_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio9_2071_2100_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO9 (i.e. average monthly mean temperature over the driest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 2071-2100, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio14_1981_2010_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio14_1981_2010_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio14_1981_2010_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO14 (precipitation of the driest month) in ".grd" and ".gri" format. Files correspond to the time-period 1981-2010, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio14_1981_2010_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio14_1981_2010_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio14_1981_2010_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO14 (precipitation of the driest month) in ".grd" and ".gri" format. Files correspond to the time-period 1981-2010, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio14_2041_2060_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio14_2041_2060_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio14_2041_2060_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO14 (precipitation of the driest month) in ".grd" and ".gri" format. Files correspond to the time-period 2041-2060, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio14_2041_2060_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio14_2041_2060_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio14_2041_2060_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO14 (precipitation of the driest month) in ".grd" and ".gri" format. Files correspond to the time-period 2041-2060, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio14_2061_2080_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio14_2061_2080_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio14_2061_2080_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO14 (precipitation of the driest month) in ".grd" and ".gri" format. Files correspond to the time-period 2061-2080, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio14_2061_2080_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio14_2061_2080_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio14_2061_2080_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO14 (precipitation of the driest month) in ".grd" and ".gri" format. Files correspond to the time-period 2061-2080, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio14_2071_2100_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio14_2071_2100_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio14_2071_2100_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO14 (precipitation of the driest month) in ".grd" and ".gri" format. Files correspond to the time-period 2071-2100, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio14_2071_2100_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio14_2071_2100_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio14_2071_2100_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO14 (precipitation of the driest month) in ".grd" and ".gri" format. Files correspond to the time-period 2071-2100, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio18_1981_2010_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio18_1981_2010_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio18_1981_2010_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO18 (average monthly precipitation during the warmest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 1981-2010, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio18_1981_2010_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio18_1981_2010_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio18_1981_2010_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO18 (average monthly precipitation during the warmest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 1981-2010, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio18_2041_2060_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio18_2041_2060_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio18_2041_2060_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO18 (average monthly precipitation during the warmest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 2041-2060, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio18_2041_2060_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio18_2041_2060_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio18_2041_2060_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO18 (average monthly precipitation during the warmest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 2041-2060, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio18_2061_2080_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio18_2061_2080_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio18_2061_2080_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO18 (average monthly precipitation during the warmest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 2061-2080, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio18_2061_2080_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio18_2061_2080_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio18_2061_2080_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO18 (average monthly precipitation during the warmest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 2061-2080, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio18_2071_2100_ssp245_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio18_2071_2100_ssp245_GFDL_CM4_ag_10.grd/.gri" "bio18_2071_2100_ssp245_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO18 (average monthly precipitation during the warmest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 2071-2100, scenario SSP2.45 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.
"bio18_2071_2100_ssp585_EC-Earth3-Veg_LR_ag_10.grd/.gri" "bio18_2071_2100_ssp585_GFDL_CM4_ag_10.grd/.gri" "bio18_2071_2100_ssp585_MPI_ESM1_2_LR_ag_10.grd/.gri"	Raster files for BIO18 (average monthly precipitation during the warmest quarter of the year) in ".grd" and ".gri" format. Files correspond to the time-period 2071-2100, scenario SSP5.85 and the climate models EC-Earth3-Veg_LR, GFDL_CM4, and MPI_ESM1_2_LR.

Spatial layers used to plot genomic offsets, define spatial masking and crop maps

Files	Description
"Countour_for_Colombia.shp"	Administrative contour layer in ".shp" format (i.e. shapefile format) representing Colombia and adjacent regions of Venezuela, used as the base map for plotting. This file was extracted and modified from IPUMS International.
"Dry_forest_merged.shp"	Polygon layer in ".shp" format (i.e. shapefile format) delimiting Tropical Dry Forest regions across Colombia and Venezuela, extracted from the Terrestrial Ecoregions of the World dataset (Olson et al. 2001). The original ecoregion spatial files are available in the ArcGIS Hub repository
"Savannas_Guiana_and_Colombia_merged_new.shp"	Polygon layer in ".shp" format (i.e. shapefile format) delimiting the savanna biome in Colombia, venezuela and other regions in northern South America, extracted from the Terrestrial Ecoregions of the World dataset (Olson et al. 2001). The original ecoregion spatial files are available in the ArcGIS Hub repository
"ECO_NAME_Magdalena-Urabá moist forests.gpkg"	Spatial layer in ".gpkg" format (i.e. GeoPackage format) delimiting the Magdalena–Urabá moist forest ecoregion extending from the Middle Magdalena Valley to the Gulf of Urabá in Colombia, extracted from the Terrestrial Ecoregions of the World dataset (Olson et al. 2001). The original ecoregion spatial files are available in the ArcGIS Hub repository

5) "Qgis_Enterolobium.zip"

"Qgis_Enterolobium.zip" is hosted by Zenodo.

This compressed .zip file contains a QGIS project file, associated spatial files as ESRI shapefiles, and attribute tables used to generate maps for Figure 1 and 2A. The files include GBIF occurrence records, biome and vegetation layers (e.g., Neotropical seasonally dry forests, moist forests, Andean montane forests, and Orinoquía savannas), country boundaries, and a population genetic structure table derived from the SNMF analysis for K=4 (the best supported number of genetic clusters). To reproduce the maps, the zip file must be extracted and the file named "Qgis_Enterolobium.qgz" opened in QGIS.

File extensions associated with spatial layers include: .shp (geometry), .shx (shape index), .dbf (attribute table), .prj/.qpj (projection information), .cpg (character encoding), and .qix (spatial index). Each file within the compressed folder is described below.

File	Description
"Qgis_Enterolobium.qgz"	QGIS project file containing map layers and associated files.
"Enterolobium_gbif_oct_2025.csv"	CSV table containing GBIF occurrence records of Enterolobium cyclocarpum.
"Enterolobium_gbif_Oct_25.dbf/.prj/.qix/.qmd/.shp/.shx"	Shapefile version of GBIF occurrence records imported into QGIS. Includes spatial coordinates and associated attribute data.
"SNMF_piecharts_Q_4_Enterolobium.csv"	Table containing ancestry coefficients (Q values for K = 4 genetic clusters) from the SNMF analysis used to generate pie-chart visualizations of genetic clusters and admixture coefficients.
"Countries_WGS84.cpg/.dbf/.prj/.shp/.shx"	Country boundary spatial layers. The file was taken from IPUMS International.
"Dry_xeric_biome.cpg/.dbf/.prj/.shp/.shx"	Shapefile representing tropical and subtropical dry/xeric regions in northern South America, created by extracting and merging Neotropical xeric ecoregions from the Terrestrial Ecoregions of the World dataset (Olson et al. 2001). The original ecoregion spatial files are available in the ArcGIS Hub repository.
"Moist_forests.cpg/.dbf/.prj/.shp/.shx"	Shapefile representing moist forest regions in northern South America, created by extracting and merging Neotropical moist forest ecoregions from the Terrestrial Ecoregions of the World dataset (Olson et al. 2001). The original ecoregion spatial files are available in the ArcGIS Hub repository.
"Montane_forests_merged_new.cpg/.dbf/.prj/.shp/.shx"	Shapefile representing montante and Andean forest regions, created by extracting and merging Neotropical montane ecoregions from the Terrestrial Ecoregions of the World dataset (Olson et al. 2001). The original ecoregion spatial files are available in the ArcGIS Hub repository.
"Savannas_Guiana_and_Colombia_merged_new.cpg/.dbf/.prj/.shp/.shx"	Shapefile representing savanna ecoregions of the Orinoquía in northern South America, created by extracting and merging the corresponding savanna ecoregions from the Terrestrial Ecoregions of the World dataset (Olson et al. 2001).The original ecoregion spatial files are available in the ArcGIS Hub repository.
"combinado_dry.cpg/.dbf/.prj/.shp/.shx"	Shapefile representing Neotropical Seasonally Dry Forest (NSDF) ecoregions, created by extracting and merging the corresponding tropical dry forest ecoregions from the Terrestrial Ecoregions of the World dataset (Olson et al. 2001). The original ecoregion spatial files are available in the ArcGIS Hub repository.
"ESRI ocean" spatial layer	This spatial layer is used as background map, and needs to be loaded after opening the "Qgis_Enterolobium.qgz" project file. The original source of the layer can be found at: https://qms.nextgis.com/geoservices/477.

Code/software

You will need R to run the scripts “Script_Genomic_Analyses_Enterolobium.R”, “SNMF_R_script.R”, and “R_script_future_climate_models.R”. For the latter, the use of a virtual machine or a computer with high RAM memory is recommended. In addition, you will need to install and load all packages specified at the beginning of each script. You will also need the software QGIS to run the files contained in “Qgis_Enterolobium.zip”. In additon, if you want to run the bash script files contained in the folder "SNP_calling_and_filtering.zip" you will need the software Stacks and VCFtools.

Notes

1) Demultiplexed ddRAD sequences used in this study are available in the ENA (European Nucleotide Archive) repository under the project with accession number PRJEB112779.

2) Results from steps 17 and 18 in “Script_Genomic_Analyses_Enterolobium.R” were stored as data frames or matrices to reduce script execution time (i.e. gene flow estimates using the function "divMigrate" in step 17, and FST estimates using the function "pairwise.WCfst" used in step 18). However, these analyses can be re-executed by removing the "#" symbol preceding the corresponding commands in the R script.

References

Ahrens, C. W., Rymer, P. D., Stow, A., Bragg, J., Dillon, S., Umbers, K. D. L., & Dudaniec, R. Y. (2018). The search for loci under selection: Trends, biases and progress. Molecular Ecology, 27(6), 1342-1356.
Döscher, R., Acosta, M., Alessandri, A., Anthoni, P., Arsouze, T., Bergman, T., Bernardello, R., Boussetta, S., Caron, L.-P., Carver, G., Castrillo, M., Catalano, F., Cvijanovic, I., Davini, P., Dekker, E., Doblas-Reyes, F. J., Docquier, D., Echevarria, P., Fladrich, U., … Zhang, Q. (2022). The EC-Earth3 Earth system model for the Coupled Model Intercomparison Project 6. Geoscientific Model Development, 15(7), 2973-3020.
Durka, W., Michalski, S. G., Höfner, J., Bucharova, A., Kolář, F., Müller, C. M., Oberprieler, C., Šemberová, K., Bauer, M., Bernt, M., Bleeker, W., Brändel, S., Bucher, S. F., Eibes, P. M., Ewald, M., Goldberg, R., Grant, K., Haider, S., Harpke, A., … Zidorn, C. (2025). Assessment of genetic diversity among seed transfer zones for multiple grassland plant species across Germany. Basic and Applied Ecology, 84, 50-60.
Held, I. M., Guo, H., Adcroft, A., Dunne, J. P., Horowitz, L. W., Krasting, J., Shevliakova, E., Winton, M., Zhao, M., Bushuk, M., Wittenberg, A. T., Wyman, B., Xiang, B., Zhang, R., Anderson, W., Balaji, V., Donner, L., Dunne, K., Durachta, J., … Zadeh, N. (2019). Structure and performance of GFDL’s CM4.0 Climate Model. Journal of Advances in Modeling Earth Systems, 11(11), 3691–3727.
Hijmans, R.(2023). raster: Geographic analysis and modeling with raster data. R package version 3.6-26.
Karger, D. N., Conrad, O., Böhner, J., Kawohl, T., Kreft, H., Soria-Auza, R. W., Zimmermann, N. E., Linder, H. P., & Kessler, M. (2017). Climatologies at high resolution for the earth’s land surface areas. Scientific Data, 4(1), 170122.
Karger, D. N., Chauvier, Y., & Zimmermann, N. E. (2023). chelsa‐cmip6 1.0: A python package to create high resolution bioclimatic variables based on CHELSA ver. 2.1 and CMIP6 data. Ecography, 2023(6), e06535.
Mauritsen, T., Bader, J., Becker, T., Behrens, J., Bittner, M., Brokopf, R., Brovkin, V., Claussen, M., Crueger, T., Esch, M., Fast, I., Fiedler, S., Fläschner, D., Gayler, V., Giorgetta, M., Goll, D. S., Haak, H., Hagemann, S., Hedemann, C., … Roeckner, E. (2019). Developments in the MPI-M Earth System Model version 1.2 (MPI-ESM1.2) and Its Response to Increasing CO2. Journal of Advances in Modeling Earth Systems, 11(4), 998-1038.
Olson, D. M., E. Dinerstein, E. D. Wikramanayake, N. D. Burgess, G. V. N. Powell, E. C. Underwood, J. A. D'Amico, I. Itoua, H. E. Strand, J. C. Morrison, C. J. Loucks, T. F. Allnutt, T. H. Ricketts, Y. Kura, J. F. Lamoreux, W. W. Wettengel, P. Hedao, and K. R. Kassem. 2001. Terrestrial ecoregions of the world: a new map of life on Earth. BioScience, 51(11), 933-938.
O’Neill, B. C., Tebaldi, C., Van Vuuren, D. P., Eyring, V., Friedlingstein, P., Hurtt, G., Knutti, R., Kriegler, E., Lamarque, J.-F., Lowe, J., Meehl, G. A., Moss, R., Riahi, K., & Sanderson, B. M. (2016). The Scenario Model Intercomparison Project (ScenarioMIP) for CMIP6. Geoscientific Model Development, 9(9), 3461–3482.
Paris, J. R., Stevens, J. R., & Catchen, J. M. (2017). Lost in parameter space: a road map for STACKS. Methods in Ecology and Evolution, 8(10), 1360–1373.
Peterson, B. K., Weber, J. N., Kay, E. H., Fisher, H. S., & Hoekstra, H. E. (2012). Double digest RADseq: An inexpensive method for de novo SNP discovery and genotyping in model and non-model species. PLoS ONE, 7(5), e37135.
Rochette, N. C., Rivera-Colón, A. G., & Catchen, J. M. (2019). Stacks 2: Analytical methods for paired-end sequencing improve RADseq-based population genomics. Molecular Ecology, 28(21), 4737-4754.
Ushey, K., Allaire, J., & Tang, Y. (2023). R package Reticulate: Interface to ‘Python’.

Genomic data

Our study deployed double digest restriction-site associated DNA (ddRAD) sequencing to generate libraries representing a subset of the genome (Peterson et al. 2012). We extracted genomic DNA using the DNA extraction Plant Kit Qiagen, double-digested the DNA and followed the library preparation protocol of Durka et al. (2025). Multiplexed samples were sequenced paired-end. After sequencing, we demultiplexed the reads using the command process_radtags from the Stacks 2.0 pipeline (Rochette et al. 2019). We recovered raw genomic sequences for 115 samples and retained 109 samples for analysis after post-processing of raw sequences. Raw DNA sequences were assembled de novo (i.e., without mapping to a reference genome, which is not available for E. cyclocarpum) using “Stacks 2” (Rochette et al. 2019). Demultiplexing was performed using the command "process_radtags" in Stacks 2.0 to perform de novo locus assembly. To optimize the de novo locus assembly and calling of SNPs, we followed established practices (Paris et al. 2017), using "m" (minimum number of raw reads required to form a stack or putative allele), "M" (allowed number of mismatches between putative alleles to define putative loci), and "n" (the allowed number of matches between putative loci during catalog construction) equal to 4, which maximized polymorphisms in the data while minimizing ‘false’ RAD loci caused by over- or under-merging. The demultiplexed ddRAD sequences used in this study are available in the ENA (European Nucleotide Archive) with accession number PRJEB112779. Following de novo locus assembly, the data were further filtered using “VCFtools” version 0.1.16 (Danecek et al. 2011) by removing indel sites (i.e., – remove-indels), allowing sites with a maximum of 50% missing data (i.e., – max-missing 0.5), including only bi-allelic sites (i.e., – min-alleles 2 and – max-alleles 2), sites with mean read depths between 5 and 100 (i.e., – min-meanDP 5, and – max-meanDP 100), and genotypes with mean read depths between 5 and 100 (i.e., – minDP 5, and – maxDP 100). Additionally, we further filtered the data by setting up the minimum percentage of individuals in a population required to process a locus equal to 20% (i.e., – r 0.2) and restricted data to only a single SNP per locus (i.e., –write-single-snp). We did not restrict sites by minimum minor allele frequencies, nor by the maximum level of heterozygosity a variable site can possess to be included, in order to recover low-frequency sites under selection (Ahrens et al. 2021). The resulting dataset comprised 109 individuals genotyped at a total of 29,266 SNPs. This dataset exhibited a 28.62% rate of missing data, which were subsequently imputed by assigning the most common genotype for each locus (represented by a single SNP) across individuals for further analyses. The file with genotype data and information on the reference and alternative alleles is provided in VCF format, as well as in PLINK format containing genotypes coded as 0 / 1 / 2 (i.e., 0, 1, or 2 copies of the alternative allele).

Future climate scenarios

We generated different climate change scenarios using the “chelsa cmip6” Python module (Karger et al. 2023), accessed via the “reticulate” R package (Ushey et al. 2023). This module provides climate projections based on the Coupled Model Intercomparison Project Phase 6 (CMIP6) (O’Neill et al. 2016), using CHELSA version 2.1 (Karger et al. 2017) as a baseline. We selected the period 1981-2010 as the baseline and projected climate data to three different periods (2041-2060, 2061-2080 and 2071-2100) under two global development scenarios (Fossil-fueled development Shared Socioeconomic Pathway 5-8.5 [SSP5-8.5] and 2-4.5 [SSP 2-4.5] emission models), and three different global climate circulation models (GFDL-CM4 [Geophysical Fluid Dynamics Laboratory`s CM4.0 physical climate model; Held et al. 2019], MPI-ESM-1.2-LR [Max Planck Institute for Meteorology Earth System Model version 1.2 - low resolution; Mauritsen et al. 2019], and EC-EARTH-3-VEG-LR [Earth Consortium Version 3 - Vegetation - low resolution; Döscher et al. 2022]). Climate data were stored as raster layers for both present and future periods, across models and scenarios, for each variable (raster files are provided in this repository). These rasters covered the main areas of distribution of the species in northern South America spanning the Caribbean, inter-Andean valleys, and the Orinoquía in Colombia and Venezuela (longitude = -87 to -67 degrees, latitude = 1 to 12.4 degrees). To improve computational efficiency, climate model rasters were downscaled by a factor of 10 by averaging groups of 10 x 10 cells using the “aggregate” function in the “raster” R package (Hijmans 2023).