A quantitative measurement can have variation, referred to here as measurement variation, which is a probability distribution. Machine Learning models typically produce a prediction corresponding to the mode of the measurement variation. The Deviation Error is a novel metric, described here, to assess predictions that accounts for measurement variation. Measurement variations in genomics data were explored. Towards a general prescription for modelling genomics measurements to reduce the Deviation Error, different loss functions were used to fit models on synthetically generated data that mimics genomics measurements. Compared to using the Mean Squared Error as the loss function, none of the other loss functions examined yielded models that performed significantly better. However, using variants of the Mean Squared Log Error and the Negative Log Likelihood as loss functions yielded models that performed significantly worse.