Competition model with GRM

[BT]

Hi,

I am trying to fit competition model with genomic relationship matrix. I get the following error 

res.comp <- remlf90(fixed = Height_m ~ TRIAL,

                    generic =list(model = c('competition'),

                    G = list(incidence = Z,precision = Ginv),

                    coord = mydata2[,c('X', 'Y')],

                    copetition_decay = 1,

                    pec = list(present = TRUE)),

                   spatial = list(model = 'AR',

                                  coord = mydata2[, c('X','Y')],

                                  rho = c(.8,.8)),

                   data = mydata2,method = 'em')

Error: Elements model, copetition_decay of the generic component must be lists.

Could you please show me how to fit the GRM with competition model.

Thank you,

BT

[Facundo Muñoz]

Hi,

The competition model is one of the "genetic" models and it works with a pedigree only.

Unfortunately, it can't be replicated using generic components because while you can have two generic components, they are assumed independent, whereas the competition model is a bi-variate joint model.

You could nevertheless do it, using a generic model for the direct genetic effect (incidence= indicator matrix for the genotypes, covariance: G matrix) and a second generic model for the indirect genetic effect (incidence= competition matrix, covariance: G matrix). It is not optimal, but you could assess the empirical correlation between the components empirically by looking at the corresponding breeding values.

Hope it helps

ƒacu.-

--
Report issues at https://github.com/famuvie/breedR/issues
---
You received this message because you are subscribed to the Google Groups "breedR" group.
To unsubscribe from this group and stop receiving emails from it, send an email to breedr+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/breedr/c6a9ae0f-b788-4763-b1e1-86a298b821aen%40googlegroups.com.

[BT]

Hi Facundo,

Thank you. I'll try as you suggested.

Regards,

BT