empiricalPvalue: Calculate the empirical P-value by affinity of cutoff.
In PWMEnrich: PWM enrichment analysis

This is the new backend function for empirical P-values for either affinity or cutoff. The function only works on single sequences.

empiricalPvalue(
  scores,
  seq.len,
  pwm.len,
  bg.fwd,
  bg.rev,
  cutoff = NULL,
  B = 10000,
  verbose = FALSE,
  exact.length = FALSE
)

`scores`	the scores obtained for the sequence
`seq.len`	the length of the sequence, if a single value will take a single sequence of given length. If a vector of values, will take sequences of given lengths and joint them together
`pwm.len`	the lengths of PWMs
`bg.fwd`	raw odds scores for the forward strand of background
`bg.rev`	raw odds scores for the reverse strand of background
`cutoff`	if not NULL, will use hit count above this cutoff. The cutoff should be specified in log2.
`B`	the number of random replicates
`verbose`	if to give verbose progress reports
`exact.length`	if to take into consideration that the actual sequence lengths differ for different PWMs. For very long sequences (i.e. seq.len >> pwm.len) this make very little difference, however the run time with exact.length is much longer.