As before, given the model parameters μ, π, we can infer the genotype at each position by modifying Equation (2) as follows: (8) The updating equations are unchanged in the M-step of EM. The model-fitting algorithm changes only in the E-step by using Equation (2) instead of Equation (2). We have specified how to encode base and mapping uncertainty into the model, obviating the need for thresholding these quantities. We call this version of the model SNVMix2.
