Having used REML to estimate the heritabilities contributed by one or more kinship matrices, you can use the results to estimate effect sizes for the predictors used to calculate the kinship matrices. These effect sizes can subsequently be used to predict phenotypes for new samples (see Calculate Scores).
Always read the screen output, which suggests arguments and estimates memory usage.
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
The main argument is --calc-blups <outfile>.
The required options are
--remlfile <remlfile> - to specify the .reml file produced by REML.
--grm <grmstem> or --mgrm <grmlist> - to specify the kinship matrices used when performing REML.
--bfile/--gen/--sp/--speed <datastem> - to specify the genetic data files used when calculating the kinship matrices (if regions were used when performing REML, these data files must also include the regional predictors).
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
The output file are
<outfile>.blup - contains the overall estimates of effect sizes for each predictor. The first four columns provide the predictor name, Allele 1 (test allele), Allele 2 (reference allele) and the predictor centre (the mean of its allele count with respect to Allele 1), with the final column providing the (raw) effect size. To work out the contribution of a predictor, subtract from its allele count the centre then multiply this by the effect size.
<outfile>.blup.full - for each predictor, this indicates how much each kinship matrix and region contributes to its overall effect size estimate.
<outfile>.pred and <outfile>.pred.full - contains the predicted values for each sample, either overall or divided into contributions from kinship matrices and regions.
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Example:
Here we use the binary PLINK files human.bed, human.bim and human.fam from the Test Datasets, and results from the example in REML.
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
First we compute effect size estimates corresponding to the results in reml1.reml (we generated this file by regressing the phenotype on the kinship matrix with stem LDAK-Thin).
./ldak.out --calc-blups blup1 --remlfile reml1.reml --grm LDAK-Thin --bfile human
blup1.blup and blup1.blup.full contain estimated effect sizes for the predictors used to calculate the kinship matrix
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Next we compute effect size estimates corresponding to the results in reml2.reml (we generated this file by regressing the phenotype on the kinship matrix with stem LDAK-Thin, plus one region).
./ldak.out --calc-blups blup2 --remlfile reml2.reml --grm LDAK-Thin --bfile human
blup2.blup and blup2.blup.full contain estimated effect sizes for the predictors used to calculate the kinship matrix and those in the region
