In the original description of SAM, the mean FDR is obtained by using the mean of the estimated number of false positives in each permutation, while the SAM software in the R package (Chu et al. 2005) provides the median FDR using the median of the estimated number of false positives in each permutation.
Figure 4 compares the runtimes of FAIS, FAISWY, BRUTE and BRUTE-WY for parameters n = 100 and J = 100 (number of permutations) while varying the sequence length L. UFE is not included because it is simply linear in L. Note that the axes are log-scaled: for L ∼ 100000, FAIS-WY takes 26.56 s, while BRUTE takes 30 min and BRUTE-WY takes ∼ 24 h.
In general, the GPD approximation usually requires fewer than 1/ Pperm permutations, while estimation using the ECDF always requires more than 1/ Pperm permutations.
Strobl et al. [86] recommended "Conditional Permutation" while calculating the variable importance to reduce the effect of spurious correlations.
Interestingly, TopSearch [ 43] could identify the exact site of circular permutation while superimposing the UBR-box with other treble clef domains.
First algorithm can solve a permutation flowshop while the other two are developed to solve a non-permutation flowshop.
GLM associations are indicated by EMP2 (genome-wide empirical value, corrected by 1000 permutations), while MLM associations are indicated by the q-value (P-value corrected by FDR < 0.05).
Such an approach can lead to a decrease in the total number of permutations, and thus computational time (or at least to a more sensible division of the total number of permutations), while producing more accurate P-value estimates.
Additionally, we can see that BEAMs power is lower than the χ test using permutation, while roughly tracks that of Bonferroni corrected χ test until the MAF=0.5 case.
A reversal is responsible for reversing the order and flipping the signs of a sequence of elements within a permutation, while a transposition is responsible for switching the location of two contiguous portions of a permutation.
