This species has relatively wide distribution and we are interested in clustering / geographic patterns of variation and genetic signals of demographic history. Any approach assuming random mating / standard Hardy-Weinberg approaches are clearly not appropriate. Data set is not overly large (48 individuals).
Hi Paul
You can try Tess3r https://github.com/cayek/TESS3/tree/master/tess3r
It uses a model free approach that could be appropriate for your data
Pablo
Try using a Discriminant Analysis of Principal Components (DAPC) implementing in the R package adegenet. It is a multivariate approach that does not assume HWP or any underlying genetic model. Like STRUCTURE, you can also get assignment probabilities. Other nice things: it easily integrates datasets the size of a typical RAD-seq project and adegenet has other functions to perform spatial analyses.
ThAlso a great suggest - was leaning towards at least trying a PCA approach. Any other PCA packages I should Coscider?
For PCAs I almost always use adegenet (actually I use another R package for the PCA part [FactoMineR] based on a genind object from adegenet, but you can follow the adegenet tutorial too). I just stumbled across this today, which is also part of adegenet ( Article A fast likelihood solution to the genetic clustering problem
). Try it out and let me know how it goes.- Badges
- Science topic