A set of methods for using wavelet transforms to characterize spatial pattern in allele frequency
Currently incomplete but in progress.
Published here in Genetics https://academic.oup.com/genetics/article/227/3/iyae082/7676102
The R scripts above contain functions for implementing these methods, along with two scripts demonstrating their use.
This file ExampleGenomicWavDissim2.R, contains a script with example SNP and geographic coordinate datasets (included in this repository). This shows how to use my code to calculate genome-wide wavelet dissimilarity, using the SNP matrix and the coordinates of samples. It also shows how to plot the mean dissimilarity among sites compared to a null, as well as a the local dissimilarity at each sampled site compared to a null, across a range of spatial scales.
SLiM simulations were conducted with linear selective gradient and a single distinct habitat patch. SpaVarSel_linear_sigma1_sigK1_JRL.1.5.slim has the representative linear gradient simulation. SpaVarSel_patchy_sigma2_sigK0.5_JRL.1.5.slim has the representative patch simulation.
An example dataset is included, resulting from a run of SpaVarSel_patchy_sigma2_sigK0.5_JRL.1.5.slim using a scale of mating and dispersal sigma = 0.5, and a strength of selection (sigma_K2) = 0.125.
The script ScalSpGenVarTest_Example_1_S.R shows how to implement the scale-specific genetic variance test on this example dataset. The script assumes the user has already filtered loci for minor allele frequency (MAF) < 0.1. Additionally, one should filter loci that show no genotypic variation (i.e. when all individuals are heterozygote) which often arises due to mapping errors.