Hi, I'm wondering if there's any documentation on how the .cns file output of CNVKit is generated from the .cnr, specifically how the .cns segments are defined. As I understand, each segment comprises .cnr bins that have similar log2 ratios, but when I plot the log2 of the bins corresponding to a particular segment (all bins with the genes listed in the 'gene' column in the .cns file), the variance is huge, from -27 to 3. The reported log2 in the .cns file is -14, which is pretty close to what I get when I take the mean log2 across the bins, but given the variance is so large, I'm wondering why these were defined as a single segment to begin with.
A related question, if I am trying to classify a particular gene's copy number category (loss, neutral, or gain), would it be best to use the log2 value from the .cns file, or would it be better to take some sort of average across the relevant bins in the .cnr file instead?
Thanks very much for the help!
