Segments do not take into account chromosome sizes

[lbeltrame]

When trying to associate an aberration with these coordinates

16 90000001 90500000

to the actual cytoband it resides in, we realized that that the bin end (and thus the segment's) goes beyond the chromosome boundary (90338345), which causes some problems with downstream tools (like UCSC's Table Browser).

I'm not sure if it's how feasible is to adjust this, but perhaps at least in the final output the coordinates should not go beyond the actual boundary of a chromosome (yes, I'm aware this opens a can of worms on how to get said sizes).