genetic distance as covariance matrix through ginverse?

I am currently using MCMCglmm to model species ability to cross with one
another. For most of the datasets species relationships can be defined by a
phylogeny and I can simply incorporate this information into the model
using the pedigree or ginverse options.

I would like to run similar models now but for subspecies/races that do not
have well defined phylogenies (the questions is identical). I have genetic
distances for some of the species in these data sets (some I have complete
pairwise genetic distances, others I only have distances for specific
species pairs). I have considered making a  simple NJ tree based on
distance matrices and incorporating this "phylogeny" via pedigree or
ginverse, but this strategy will not work without complete pairwise
distances.

Is there a way to utilize the distance matrix (complete or more sparse)
directly? Since the genetic distance matrix looks very similar to the
relationship matrix "A" in the MCMCglmm tutorial and coursenotes I tried
incorporating it into the model but have run into a few errors.

Here is the fake data I have been playing around with. Any help would be
greatly appreciated

#Genetic Distance Matrix. I have also been replacing NA with 0 but still
encounter same error

x <- c( 0.07 ,0.08, 0.11,NA,

  0.16, NA)

M <- matrix(0, 4, 4)

M[lower.tri(M)] <- x

M <- t(M)

M[lower.tri(M)] <- x

dimnames(M) <- list(c("sp1","sp2","sp3","sp4"),c("sp1","sp2","sp3","sp4"))

diag(M)<-1

#Fake response variable, species ID, and categorical predictor


y<-c(5,5,6,6,10,10,12,11,4,3,4,5,11,12,13,14)

species<-c("sp1","sp1","sp1","sp1","sp2","sp2","sp2","sp2","sp3","sp3","sp3"
,"sp3","sp4","sp4","sp4","sp4")

species<-as.factor(species)

ms<-c(0,0,0,0,1,1,1,1,0,0,0,0,1,1,1,1)

data<-data.frame(y,ms,species)


# want to run a model for y~ma taking into account species

library(MCMCglmm)

#define prior

p.var<-var(y)

prior1=list(R=list(V=p.var/2, n=1), G=list(G1=list(V=p.var/2, n=1)))

library(MASS)

M1<-ginv(M) #I have also been using M in the model with the same error

dimnames(M1) <- list(c("sp1","sp2","sp3","sp4"),c("sp1","sp2","sp3","sp4"))

model1<-MCMCglmm(y~ms,random=~species,ginverse=list(species=M1),prior=prior1
,data=data)

#The error I receive.

Error in FUN(X[[1L]], ...) :

   trying to get slot "Dim" from an object of a basic class ("matrix") with
no slots

      [[alternative HTML version deleted]]