Fig. 3From: Approximate Bayesian neural networks in genomic predictionModel-averaged weight plots for the dominance genetic effects for the best model (one-neuron, one-layer and linear activation function) on the simulated QTLMAS2010 data. The dominance, over-dominance and under-dominance loci were identified at SNPs 9212, 9404 and 9602, respectively (blue circles)Back to article page