I have run superFreq on a couple of samples from gallbladder tissues along with OCUG and NOZ cell-line which have ERBB2 amplifications. The combined CNA heatmap that lists out the calls gene-wise can call the amplifications in these samples, but the tsv file stating the M statistic, defined as,

"Consensus Log2 fold change of read depth with respect to the reference normals across the segment, normalized for ploidy so that 0 corresponds to 2 copies, -1 corresponds to 1 copy, +1 is four copies, etc."

is negative for the above cell lines, with OCUG-1 having a value of -0.815051998801712 and NOZ with -1.92131833157471 (Also stating the number in tabular form).

             M statistic (superFreq)
OCUG-1          -0.815051998801712
NOZ         -1.92131833157471