Wethen simulate k+1 components by constructing a g??(k+1) matrix Ctrueoffundamental components in which each column is an end-member (i.e. sub-type) and each row is the copy number of one hypothetical gene, sampledfrom the unit Gaussian distribution and rounded to the nearest integer.Samples mi, corresponding to the columns of the data matrix M, are thengiven by:mi=2log2where ?????normal(0,1) and