REML is used for estimating the proportions of phenotypic variance explained by kinship matrices (their heritability contributions). The variance explained by a kinship matrix represents the variance explained by the predictors across which the kinships are computed. Although traditionally kinship matrices represent genome-wide correlations and are calculated across many tens of thousands of SNPs, it is equally valid to compute kinships across far fewer SNPs, for example, if wishing to consider the variance explained by a gene or small genomic region. However, in this case, instead of first explicitly calculating the kinship matrix across this region, then performing REML, it is usually much more efficient to do this on-the-fly as part of the REML algorithm.
The generalized REML solver contained in LDAK is optimized for the case when some kinship matrices correspond to small subsets of SNPs (referred to as regions). Also, when the phenotype is binary, it can convert estimates of variance explained from the observed scale to the liability scale.
If your aim is prediction, then having estimated the variance components, you can then use the BLUP feature to estimate SNP effect sizes. These effect size estimates can subsequently be used for predicting the phenotypic values of new individuals given their genotypes.