This method looks interesting, although in the study presenting it there was little overlap with other classical methods such as LFMM or BayEnv. An interesting feature of this software is that it outputs information about spatial autocorrelation of genotypes (using Moran’s I) to correct for spurious association. Population structure can also be included (multivariate analysis) to correct for relatedness between samples. Also, the software allows for converting data to a suitable format from PLINK input files (this private format is something that would have been nice to avoid but I am picky). Having read the paper, I would suggest to use it in combination with other landscape genomics programs (LFMM, BayPass…) and look at common candidates. It makes a nice addition to the landscape genomics toolbox.
You can check the paper here (open access): http://onlinelibrary.wiley.com/doi/10.1111/1755-0998.12629/full
And thanks to Anne Roulin for pointing this software to me 🙂