[R] p-value from fisher.test odds ratio 'estimate' (without 2x2 table)

[dschruth]

Hello,

I'm trying to generate a p.value directly from the output of a fisher
exact test odds ratio statistic without resorting to performing
another test on the 2x2 table itself.   The motivation for this stems
from a need to modify the test statistic before calculating it's new p-
value.

I've looked at the c source code for both fisher.test and phyper and
they both seem to operate directly on the 2x2 table rather than the
statistic.  Can anybody think of a way to do this short of recovering
the cell probabilities from the odds ratio (as described here:
http://en.wikipedia.org/wiki/Odds_ratio#Recovering_the_cell_probabilities_from_the_odds_ratio_and_marginal_probabilities)?

Thank you,

Dave