Update hmm.py, make the window paramter activate of segment_hmm function

[zhuying412]

make the window paramter activate of segment_hmm function
#833

[etal]

Thanks, @Zhu-ying . Was this change enough to fix the false negative you saw in #833 ? Did it improve your results overall, and not cause any other false positives or negatives elsewhere?

[zhuying412]

Thanks, @Zhu-ying . Was this change enough to fix the false negative you saw in #833 ? Did it improve your results overall, and not cause any other false positives or negatives elsewhere?

Yes, if I set the window=2, I can get more CNVs which is smaller size，and the false negative CNV can be called. the window
is larger, the sensitivity is lower. I don't know why, but it works. I think it will produce some false positives, but we need lower false negatives in clincal. we just want to make the -t (threshold/window) parameter actived，now the code is commented.

[etal]

Great, thanks for your contribution!