log-odds score cutoff when dumping PWMs to a file

[mbstadler]

A PWM match-score cutoff is mandatory when dumping PWMs into a file in Homer format.

Currently, the score cutoff is constant (log(2^10)) and identical for all PWMs.

Maybe the function should accept an argument to allow to choose between a constant cutoff (current behaviour), or a cut-off that is min(constant, PWMmaxScore), or a relative cut-off (e.g. 80% of PWMmaxScore).

[mbstadler]

In addition, the conversion between PWMList and homerfile should be implemented in such a way that motif names will be consistent (same SE rownames returned by calcBinnedMotifEnr() for method = "R" and method = "Homer".