Fig. 1From: Interpreting single-step genomic evaluation as a neural network of three layers: pedigree, genotypes, and phenotypesFramework of single-step NNMM with three fully-connected sequential layers of data: pedigree, genotypes, and phenotypes. Between the layer of pedigree and the layer of genotypes, the gene content of each marker is treated as a quantitative trait, and the pedigree is used to define the random effects covariance matrix. Each node in the middle layer represents the gene content of one marker. “NA” denotes missing values. For example, the nodes in the middle layer may be 2,2,0,1,0 for a genotyped individual or all missing (“NA”) for a non-genotyped individual. For non-genotyped individuals, all gene contents are missing and will be sampled conditional on pedigree, genotypes, and phenotypes in MCMCBack to article page