Testing sequence depth thresholds for assignment
To test whether lower coverage data would affect our ability to accurately assign breeding individuals, we used the 47 individuals from the breeding populations, that were not used in the ESSBPs, as a testing set for further down sampling. We did not down sample individuals from the ESSBPs because we did not want to lower the effective sample size of the source populations, but rather test how well we could assign individuals with lower coverage given the effective sample sizes of our source populations (and the amount of genetic differentiation among them). The testing set consisted of 22 individuals from the Southern Temperate population, 14 individuals from the Northern Temperate population, 6 individuals from Basin Rockies population, and 5 individuals from the Western Boreal population. The sequencing depth of these 47 individuals ranged from 0.6X – 2.0X. Down sampling these individuals to 0.1X resulted in 100% assignment accuracy, and further down sampling to 0.01X resulted in 97.9% accuracy (one individual from the Southern Temperate population assigned to the Northern Temperate population; Table 1, Table S4). The individual incorrectly assigned from the Southern Temperate population was from a sampling site in Pennsylvania which is on the border of our boundary for the Southern Temperate and Northern Temperate populations.