The figure above shows correspondence between the true genetic value (inner square ordering) and different model components under varying levels of spatial adjustment in Replicate 1 of Map 1 from a new study in The Plant Genome that tested several methods to separate genetic signals from field variability using real soybean data and computer simulations. (a): No adjustments, raw phenotypic information. (b): Model incorporates line and block effects. (c): Model incorporates line, block, row, and column effects. (d): Model incorporates line, block, row, column, and row-by-column effects. Phenotype (green), field variation (orange), estimated genetic signal (blue), and predicted genetic effects (gray) for four different manners to approximate genotype effect (L).