Confusion Regarding Determination of Means and Variances in Gaussian Mixture Model for Copy Number Inference

[bioPG]

Hello, I carefully read through your SMN tool source code and I think your tool is fantastic. I can understand most of it, but I feel a bit confused about how the means and variances of the 11 distributions in the one-dimensional Gaussian mixture model used to convert normalized depth into copy numbers are determined. Are they based on training with a specific dataset?