解释 p.adjust 函数的结果答案

【问题标题】：Interpreting results from the p.adjust function解释 p.adjust 函数的结果
【发布时间】：2020-08-12 09:11:45
【问题描述】：

我已使用以下代码对“ordered_pvs”730x1 向量执行了 p.adjust 函数：

ordered_pvs$Bonferroni = p.adjust(ordered_pvs$Raw_PVs, method = "bonferroni")
ordered_pvs$BH = p.adjust(ordered_pvs$Raw_PVs, method = "BH")
ordered_pvs$Holm = p.adjust(ordered_pvs$Raw_PVs, method = "holm")
ordered_pvs$Hochberg = p.adjust(ordered_pvs$Raw_PVs, method = "hochberg")
ordered_pvs$Hommel = p.adjust(ordered_pvs$Raw_PVs, method = "hommel")
ordered_pvs$BY = p.adjust(ordered_pvs$Raw_PVs, method = "BY")

下面是我从这个函数中得到的结果。我想知道我如何解释这些结果？例如，第 73 位的 BH = 1 和第 75 位的 HOLM = 1 是什么意思。我将如何解释其他结果？

我一直在努力寻找任何研究论文等来帮助我解决这个问题。

Raw_PVs rank critical_value Bonferroni           BH     Holm Hochberg    Hommel           BY

0.000000   31       0.002123    0.00000 0.000000e+00 0.000000 0.000000 0.0000000 0.0000000000
0.000000   32       0.002192    0.00000 0.000000e+00 0.000000 0.000000 0.0000000 0.0000000000
0.000001   33       0.002260    0.00073 2.147059e-05 0.000698 0.000697 0.0006950 0.0001539644
0.000001   34       0.002329    0.00073 2.147059e-05 0.000698 0.000697 0.0006950 0.0001539644
0.000002   35       0.002397    0.00146 4.171429e-05 0.001392 0.001392 0.0013800 0.0002991308
0.000003   36       0.002466    0.00219 5.763158e-05 0.002085 0.002079 0.0020670 0.0004132729
0.000003   37       0.002534    0.00219 5.763158e-05 0.002085 0.002079 0.0020670 0.0004132729
0.000003   38       0.002603    0.00219 5.763158e-05 0.002085 0.002079 0.0020670 0.0004132729
0.000006   39       0.002671    0.00438 1.095000e-04 0.004152 0.004146 0.0041160 0.0007852185
0.000006   40       0.002740    0.00438 1.095000e-04 0.004152 0.004146 0.0041160 0.0007852185
0.000007   41       0.002808    0.00511 1.216667e-04 0.004830 0.004823 0.0047950 0.0008724650
0.000007   42       0.002877    0.00511 1.216667e-04 0.004830 0.004823 0.0047950 0.0008724650
0.000009   43       0.002945    0.00657 1.493182e-04 0.006192 0.006183 0.0061560 0.0010707525
0.000009   44       0.003014    0.00657 1.493182e-04 0.006192 0.006183 0.0061560 0.0010707525
0.000013   45       0.003082    0.00949 2.108889e-04 0.008918 0.008918 0.0088790 0.0015122726
0.000017   46       0.003151    0.01241 2.697826e-04 0.011645 0.011645 0.0116110 0.0019345962
0.000020   47       0.003219    0.01460 3.106383e-04 0.013680 0.013680 0.0136600 0.0022275701
0.000024   48       0.003288    0.01752 3.650000e-04 0.016392 0.016392 0.0163680 0.0026173949
0.000060   49       0.003356    0.04380 8.938776e-04 0.040920 0.040920 0.0406800 0.0064099467
0.000077   50       0.003425    0.05621 1.124200e-03 0.052437 0.052437 0.0518980 0.0080615763
0.000089   51       0.003493    0.06497 1.273922e-03 0.060520 0.060520 0.0597190 0.0091352215
0.000148   52       0.003562    0.10804 2.077692e-03 0.100492 0.100492 0.0984200 0.0148990172
0.000162   53       0.003630    0.11826 2.231321e-03 0.109836 0.109836 0.1075680 0.0160006784
0.000289   54       0.003699    0.21097 3.906852e-03 0.195653 0.195653 0.1907400 0.0280158197
0.000305   55       0.003767    0.22265 4.048182e-03 0.206180 0.206180 0.2003850 0.0290292891
0.000315   56       0.003836    0.22995 4.106250e-03 0.212625 0.212625 0.2063250 0.0294456928
0.000338   57       0.003904    0.24674 4.328772e-03 0.227812 0.227812 0.2197000 0.0310413853
0.000524   58       0.003973    0.38252 6.595172e-03 0.352652 0.352652 0.3285480 0.0472936185
0.000686   59       0.004041    0.50078 8.431500e-03 0.460992 0.460992 0.4198320 0.0604618225
0.000693   60       0.004110    0.50589 8.431500e-03 0.465003 0.465003 0.4234230 0.0604618225
0.000711   61       0.004178    0.51903 8.508689e-03 0.476370 0.476370 0.4337100 0.0610153372
0.000731   62       0.004247    0.53363 8.606935e-03 0.489039 0.489039 0.4444480 0.0617198608
0.000767   63       0.004315    0.55991 8.887460e-03 0.512356 0.512356 0.4625010 0.0637314889
0.000780   64       0.004384    0.56940 8.896875e-03 0.520260 0.520260 0.4687800 0.0637990011
0.000807   65       0.004452    0.58911 9.063231e-03 0.537462 0.537462 0.4825860 0.0649919291
0.000864   66       0.004521    0.63072 9.556364e-03 0.574560 0.574560 0.5106240 0.0685281578
0.000931   67       0.004589    0.67963 1.014373e-02 0.618184 0.618184 0.5427730 0.0727401393
0.001001   68       0.004658    0.73073 1.074603e-02 0.663663 0.663663 0.5775770 0.0770591856
0.001099   69       0.004726    0.80227 1.162710e-02 0.727538 0.727538 0.6253310 0.0833773047
0.001229   70       0.004795    0.89717 1.281671e-02 0.812369 0.812369 0.6747210 0.0919079529
0.001296   71       0.004863    0.94608 1.330222e-02 0.855360 0.855360 0.6998400 0.0953895036
0.001312   72       0.004932    0.95776 1.330222e-02 0.864608 0.864608 0.7058560 0.0953895036
0.001393   73       0.005000    1.00000 1.392919e-02 0.916594 0.916594 0.7355040 0.0998854492
0.001412   74       0.005068    1.00000 1.392919e-02 0.927684 0.927684 0.7427120 0.0998854492
0.001678   75       0.005137    1.00000 1.633253e-02 1.000000 0.999309 0.8282769 0.1171196978
0.002392   76       0.005205    1.00000 2.297579e-02 1.000000 0.999309 0.9663680 0.1647581221
0.002783   77       0.005274    1.00000 2.607410e-02 1.000000 0.999309 0.9822422 0.1869759547
0.002786   78       0.005342    1.00000 2.607410e-02 1.000000 0.999309 0.9822422 0.1869759547
0.002966   79       0.005411    1.00000 2.740734e-02 1.000000 0.999309 0.9861329 0.1965365397
0.003544   80       0.005479    1.00000 3.233900e-02 1.000000 0.999309 0.9899314 0.2319011895
0.004110   81       0.005548    1.00000 3.704074e-02 1.000000 0.999309 0.9916178 0.2656171137
0.004585   82       0.005616    1.00000 4.081768e-02 1.000000 0.999309 0.9930539 0.2927013582
0.006034   83       0.005685    1.00000 5.307012e-02 1.000000 0.999309 0.9973494 0.3805629137
0.006139   84       0.005753    1.00000 5.335083e-02 1.000000 0.999309 0.9976873 0.3825758902
0.007263   85       0.005822    1.00000 6.237635e-02 1.000000 0.999309 0.9993090 0.4472973946
0.007686   86       0.005890    1.00000 6.524163e-02 1.000000 0.999309 0.9993090 0.4678441237
0.007949   87       0.005959    1.00000 6.669851e-02 1.000000 0.999309 0.9993090 0.4782913146
0.008247   88       0.006027    1.00000 6.772596e-02 1.000000 0.999309 0.9993090 0.4856590970
0.008257   89       0.006096    1.00000 6.772596e-02 1.000000 0.999309 0.9993090 0.4856590970
0.008733   90       0.006164    1.00000 7.083433e-02 1.000000 0.999309 0.9993090 0.5079491066
0.008877   91       0.006233    1.00000 7.089728e-02 1.000000 0.999309 0.9993090 0.5084005124
0.008935   92       0.006301    1.00000 7.089728e-02 1.000000 0.999309 0.9993090 0.5084005124
0.009793   93       0.006370    1.00000 7.686978e-02 1.000000 0.999309 0.9993090 0.5512289980
0.009972   94       0.006438    1.00000 7.713411e-02 1.000000 0.999309 0.9993090 0.5531244245
0.010038   95       0.006507    1.00000 7.713411e-02 1.000000 0.999309 0.9993090 0.5531244245
0.010260   96       0.006575    1.00000 7.801875e-02 1.000000 0.999309 0.9993090 0.5594681632
0.010469   97       0.006644    1.00000 7.878732e-02 1.000000 0.999309 0.9993090 0.5649795334
0.010678   98       0.006712    1.00000 7.954020e-02 1.000000 0.999309 0.9993090 0.5703784267
0.010962   99       0.006781    1.00000 8.083091e-02 1.000000 0.999309 0.9993090 0.5796340013
0.011441  100       0.006849    1.00000 8.198462e-02 1.000000 0.999309 0.9993090 0.5879071657
0.011445  101       0.006918    1.00000 8.198462e-02 1.000000 0.999309 0.9993090 0.5879071657
0.011564  102       0.006986    1.00000 8.198462e-02 1.000000 0.999309 0.9993090 0.5879071657
0.011610  103       0.007055    1.00000 8.198462e-02 1.000000 0.999309 0.9993090 0.5879071657
0.011680  104       0.007123    1.00000 8.198462e-02 1.000000 0.999309 0.9993090 0.5879071657
0.012043  105       0.007192    1.00000 8.372752e-02 1.000000 0.999309 0.9993090 0.6004054661
0.012260  106       0.007260    1.00000 8.443208e-02 1.000000 0.999309 0.9993090 0.6054577673
0.012495  107       0.007329    1.00000 8.524626e-02 1.000000 0.999309 0.9993090 0.6112962518
0.012959  108       0.007397    1.00000 8.649173e-02 1.000000 0.999309 0.9993090 0.6202274170
0.012984  109       0.007466    1.00000 8.649173e-02 1.000000 0.999309 0.9993090 0.6202274170
0.013033  110       0.007534    1.00000 8.649173e-02 1.000000 0.999309 0.9993090 0.6202274170
0.013176  111       0.007603    1.00000 8.665297e-02 1.000000 0.999309 0.9993090 0.6213837010

非常感谢！

【问题讨论】：

标签： r statistics

【解决方案1】：

如果您阅读手册：

前四种方法旨在对家庭方面的错误率。

如果方法在c("holm", "hochberg", "hommel", "bonferroni") 中，则调整后的 p 值是调整后的 alpha（显着性阈值），此条目将显着。举个例子，

p = c(0.001,0.005,0.01,0.22,0.33,0.44,0.55,0.66,0.77,0.88)

具有 alpha（显着性水平）= 0.05 的 Bonferroni 校正意味着您需要小于 0.05/10 = 0.005 才能使其显着，这意味着您将犯一个或多个 I 类错误的概率控制在 0.05。

如果我们使用 p.adjust，您会看到只有第一个在 alpha 为 0.05 时是显着的：

p.adjust(p,"bonferroni")
 [1] 0.01 0.05 0.10 1.00 1.00 1.00 1.00 1.00 1.00 1.00

对于其他 FDR 方法，这些反映了条目将通过的最低 FDR 截止值，例如：

 p.adjust(p,"BH")
 [1] 0.01000000 0.02500000 0.03333333 0.55000000 0.66000000 0.73333333
 [7] 0.78571429 0.82500000 0.85555556 0.88000000

因此，如果我们将 FDR 截止值设置为 0.05，这意味着我们允许 5% 的 I 类错误，那么我们将声明前三个条目在 5% FDR 处显着，因为前三个条目

【讨论】：