CANCOR
The CANCOR test was based on scripts by (Blanco-Pastor et al., 2021). The CANCOR test was used to analyse both environmental and phenotypic data to find loci responsible for adaptations. Environmental and phenotypic data were scaled using scale() function in R and then run against the genotypic data using the CCorA function in the R package vegan [34]. Then the significance of outlier loci was tested using an χ2 test on Mahalanobis distances. Loci were considered an outlier if the FDR=0.1. The best represented loci in the first two canonical dimensions were selected. To simplify results we removed soil parameters as their projection norms weren’t above 0.9 for any soil parameter, meaning few significant loci could be found. Loci with an alternative allele frequency that was relativly high (|r |>0.25) were kept. This resulted in lists of loci that were positively or negatively correlated with a trait of environmental parameter.