On the use of resampling tests for evaluating statistical significance of binding-site co-occurrence.

Independent resampling A major reason for the highly significant p-values observed with the permutation test approach is that the observed score is maximal and there are so many ways in which m sites of each factor can be arranged to result in fewer overlaps.

also in 132,142 other papers

We are not entirely at ease with treating one of a pair of factors differently from the other nor with the possibility that a co-occurrence may be significant when one factor is fixed but not when the other is fixed.

also in 13,097 other papers