Skip to main content
Dryad

Abies religiosa (sacred fir) genetic data from common garden seedlings (Mexico), comparing initial planting to the small subset of seedlings that survived in a low-elevation (hot, dry) common garden

Moran, Emily1; Du, Yinyue1; Cruzado-Vargas, Ana Laura23; Sáenz-Romero, Cuauhtémoc23; Zhang, Ruli4

Research facility: University of California, Davis
Published Nov 05, 2026 on Dryad. https://doi.org/10.5061/dryad.8gtht7726

Data files

Abstract

The suitable range of sacred fir (Abies religiosa), the only roosting tree for the monarch butterfly (Danaus plexippus) populations wintering in Mexico, is expected to shrink due to the impact of climate change. A common garden experiment carried out within the Monarch Butterfly Biosphere Reserve (MBBR) in 2019 - 2022  showed high >95% seedling survival across source populations at 3400 m and 3000 m (representing the upper and middle elevations of the current range), but only 6% survival at 2600 m (below the natural range, representing warmer future conditions). Of the 11 planted populations, lower elevation sources had slightly higher survival in the low elevation garden, but some individuals from high elevation populations survived as well.

We tested whether these low-elevation survivors were genetically different from the starting population (represented by the higher elevation gardens where almost all individuals survived) using genotyping by sequencing. Needle samples were dried over silica gel and shipped to UC Merced. DNA was extracted using the Qiagen DNeasy Plant Pro Kit, and purified DNA was sent to the UC Davis Genome Center for GBS library preparation with ApeKI and sequencing on Illumina HiSeq 4000.

This dataset is output from the STACKS-2.59 de novo pipeline, including haplotypes for each individual, SNP calling output, consensus sequence for each locus, Fst comparisons, allele frequencies, etc. Reads were demultiplexed, cleaned, and trimmed to 64 bp. The quality score limit was set to 20, window size to 0.20, and other parameters were set to default settings. To limit false SNP identification and increase accuracy of downstream analyses, SNPs were filtered using VCFtools. We selected only biallelic SNPs (min-alleles 2, max-alleles 2), where at least 25% of individuals were genotyped at that SNP (max-missing 0.75), and the minor allele frequency was at least 5% (maf 0.05), minDP 5, hwe 0.00001.

We used these data to conduct Principal component analysis (PCA) for all 18,816 SNPs, and population genetic differences were calculated using the population genetics program in STACKS-2.59 for initial and survivor groups. Observed heterozygosity, expected heterozygosity, observed homozygosity, expected homozygosity, an estimate of nucleotide diversity, and the inbreeding coefficient of an individual (I) relative to the subpopulation (S) were automatically calculated by STACKS. Initial and survivor groups were pairwise compared to compute the F~ST ~for each locus.