pairwise.WCfst(genind2hierfstat(x)) and genet.dist(x, method = "WC84") do not yield the same result

[DaniEstevezBarcia]

Hi,

I have just realized that these two functions do not give the same result when they both should be providing (according to the descriptions) pairwise estimated of Fst following Weir and Cockerham 1984.

Best regards

[jgx65]

I'm puzzled, it does for me? (see below). Could you provide an example?

> library(hierfstat)
> data(gtrunchier)
> pairwise.WCfst(gtrunchier[,-2])
          1         2         3         4
1        NA 0.4695425 0.4194788 0.4486061
2 0.4695425        NA 0.7795719 0.7648904
3 0.4194788 0.7795719        NA 0.6321912
4 0.4486061 0.7648904 0.6321912        NA
5 0.3356973 0.5994008 0.5688411 0.5199231
6 0.3992087 0.7563057 0.5308519 0.4776794
          5         6
1 0.3356973 0.3992087
2 0.5994008 0.7563057
3 0.5688411 0.5308519
4 0.5199231 0.4776794
5        NA 0.5065197
6 0.5065197        NA

> genet.dist(gtrunchier[,-2],method="WC84")
          1         2         3         4         5
2 0.4695425                                        
3 0.4194788 0.7795719                              
4 0.4486061 0.7648904 0.6321912                    
5 0.3356973 0.5994008 0.5688411 0.5199231          
6 0.3992087 0.7563057 0.5308519 0.4776794 0.5065197

[DaniEstevezBarcia]

do you add the [, -2] to remove the population column?

[DaniEstevezBarcia]

Also, in the genet.dist option I used a genind object. When changed to hierfstat it works, but I thought it also did with genind (it does actually provide a result but it is different)

[jgx65]

gtuncatula contains 2 populations column (hierarchical design), the [,-2] is just to remove one of them. With nancycats, the following give me TRUE:

all.equal(mat2vec(pairwise.WCfst(genind2hierfstat(nancycats))),mat2vec(as.matrix(genet.dist(nancycats,method="WC84"))))

[jgx65]

Note that if you update hierfstat with the current version on github, you won't need the extra genind2hierfstat in pairwise.WCfst (see https://github.com/jgx65/hierfstat/issues/56)

[DaniEstevezBarcia]

The new version is not on cran yet right?

Thank you for the answers

[jgx65]

no, on github

[DaniEstevezBarcia]

Ok, the problem was that the version on cran for some reason was doing something odd with the genind (it did not return error though). With the github version it works.