Can you provide some sample data and expected results?

Update: First pass through the code looks okay to me (without looking at the shuffle module) and I think I know what the input looks like: an Id line followed by a variable length string of space-separated tokens of DNA groups with possibly very long lines. So are your differing results a format difference (I suspect not) or sequence difference? For random shuffling you can't expect the same results unless you use the same random number sequence as in the comparative study.

Update: I've tried the code and, assuming data like this,

>ID1
A/B/C/D  E/F/G/Z
[download]

there are format errors in the output. We really need more information about the problem, especially the exact input format. (See toolic's comments below.)