Wild avocado (Persea americana) herbarium genomes suggest alternate path to domestication

We sequenced at 1-5x coverage (except for 3 individuals, which were 0.075-0.22x) the genomes of 25 putatively wild herbarium avocado leaves collected in the last 60 years and spanning their entire native geographic range. Illumina library preparation followed the Kapa HiFi MasterPrep Kit using Truseq indexes. AdapterRemoval trimmed adapter sequences and merged paired-end reads, and the collapsed reads were mapped to the reference genome. We used bioinformatic analyses that examine genotype likelihoods to compare and contrast the population structure of our wild avocados with that of a previously published cultivar dataset. Wild avocados are most likely structured in two distinct populations, one in Central Mexico and one spanning from Chiapas to as far as Peru, and we predict the valley between the Sierra Madre del Sur and the Sierra Madre de Chiapas acts as a reproductive barrier. Overall, wild avocado populations are more genetically differentiated and are more diverse within compared to cultivars. We attribute this disparity to the domestication process which acts to erode genetic variation over time and then reduce differences between varieties after commercial hybridization. Based on which herbarium specimens have higher genetic affinity to any of the three horticultural races, we support claims that each variety has distinct and separate domestication origin throughout Central America. We also offer a new model fitting our data that includes a single domestication event in Honduras that gives rise to both the Guatemalan and Lowland varieties.