The most basic use of MultiBLUP is when we have already computed a number of kinship matrices (which because the aim is prediction we advise doing with weightings turned off) and wish to perform generalised BLUP. First we estimate the variance components.

--reml <output> --pheno <phenofile> --mgrm <grmlist>

where grmlist provides the stems of the pre-computed kinship matrices and <phenofile> provides the phenotypic values (stored in PLINK format). The variance estimates will be saved in <output>.reml, with the random effects in <output>.indi.blp.

To include covariates add the option --covar. Use --keep to consider only a subset of samples; this is useful if wishing to fit the model on a training sample and then measure predictive accuracy for a test sample.

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Next we obtain estimates of predictor effect sizes.

--calc-blups <output> --mgrm <grmlist> --remlfile <output>.reml --bfile test

This step requires one of –bfile/–chiamo/–sp/–speed <prefix> - to provide the data files from which the kinship matrices were calculated (see File Formats).

The SNP effect sizes will be saved in <output>.blup, with the random effects in <output>.pred.

If <grmlist> contains only one kinship matrix (or if instead of --mgrm you use --grm), then the above commands will be equivalent to running (standard) BLUP.

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

The following example uses the files provided in the Test Datasets. The kinship matrices kins/kinshipA and kins/kinshipB were created from the binary PLINK files test.bed, test.bim and test.fam. Phenotypes are stored in phen.pheno and the file mlist.txt contains

kins/kinshipA

kins/kinshipB

The commands

../ldak.out --reml out --mgrm mlist.txt --pheno phen.pheno

../ldak.out --calc-blups out --mgrm mlist.txt --remlfile out.reml --bfile test

will perform 2-way MultiBLUP, saving the effect size estimates in out.blup and the predictions in out.pred. Column 4 of out.blup will report cumulative effect sizes, which are the sum of the effect sizes corresponding to each kinship matrix (Columns 6 and 8). Similarly, Column 3 of out.pred will report cumulative predictions, which are the sum of predictions corresponding to each kinship matrix (Columns 5 and 7).