An object of class `dmmblockarray`

is a list containing two items called `array`

and `blocks`

. Item `array`

is an array of objects of type `dmm`

, with each array element representing the reult of a `dmm()`

analysis for one pair of blocks of traits. The function `condense.dmmarray`

will recombine these results into a single object of class `dmm`

with the variance component and genetic parameter estimates matrices being for all traits, and other elements of the \ codedmm object being appropriately pooled. Item `blocks`

is a list, each element of which is a list of the traits present in each block.

1 |

`da` |
An object of class |

In bringing together the results of several `traitsblockwise`

analyses into a single matrix of (for example) individual additive genetic variance/covariance components, one is putting together into one matrix elements estimated with different precisions due to differnt replication for each block of traits. The resulting matrix may not be positive definite even if all the contributing `traitsblockwise`

ni x nj matrices are forced positive definite.

If the argument `da`

contains results from a `dmm`

run with `gls=T`

then the GLS results for each traitblockpair will also be condensed. In this case the GLS results must be present for every traitblockpair. It can be quite difficult to get `gls=T`

runs to converge successfully for every pair of blocks.

An object of class `dmm`

, containing the recombined results for all traits.

An object of class `dmmblockarray`

can be manipulated as is without using the `condense.dmmblockarray`

function. For example one element of the array can be printed with
print(objectname$array[[i,j]])

where `i`

and `j`

are subscripts indicating the row and column position of the element to be printed. The `summary()`

and `gsummary()`

functions are also available. It is necessary that the `library(dmm)`

statement be made, otherwise one will get the standard `print()`

and `summary()`

functions instead of those appropriate for an object of class `dmm`

. The double square brackets are necessary because each element of the array is a list object, and you want its contents, not its name attribute.

Neville Jackson

Functions `dmm()`

, `condense.dmmarray()`

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 | ```
library(dmm)
# prepare the dataset sheep.df
data(sheep.df)
# add a matrix 'Ymat' to the dataframe, which is required for traitsblockwise
# keep=TRUE also required
sheep.mdf <- mdf(sheep.df,pedcols=c(1:3),factorcols=c(4:6),ycols=c(7:9),
sexcode=c("M","F"),keep=TRUE)
# make sheep.fit as a class dmmarray object
sheep.fit <- dmm(sheep.mdf, Ymat ~ 1 + Year + Sex,traitsblockwise=TRUE,
Block1=c("Cww","Diam"),Block2="Bwt")
# look at one element of the dmmblockarray
summary(sheep.fit$array[["Block1","Block2"]])
# condense the dmmblockarray to a class dmm object
sheep.condense <- condense.dmmblockarray(sheep.fit)
# compute a response to selection
sheep.resp <- gresponse(sheep.condense,psd=list(dp=c(1,1,1)),effects=c("G(Ia)"))
# look at response object
summary(sheep.resp)
#cleanup
rm(sheep.df)
rm(sheep.mdf)
rm(sheep.fit)
rm(sheep.condense)
rm(sheep.resp)
``` |

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.

Please suggest features or report bugs with the GitHub issue tracker.

All documentation is copyright its authors; we didn't write any of that.