Skip to main content

Table 5 Simulated data: Comparison of estimation of the number of non-redundant variates from the permutation method (ENT, obtained as the average of the ENT estimates for all the simulated uncorrelated and correlated outcomes measures considered via the multivariate Normal and the multivariate log-Normal methods) and via the approximation procedure based on the eigenvalues of the correlation matrix of the metabolite concentrations (Meff)

From: Multiple-testing correction in metabolome-wide association studies

Correlation \(\in\) [0.95-1) [0.85,0.95) [0.75,0.85) [0.65,0.75) [0.55,0.65) [0.45,0.55) [0.35,0.45) [0.25,0.35) (0,0.25)
ENT 3(0.5%) 11(1.7%) 36(5.4%) 81(12.4%) 134(20.5%) 210(32%) 319(48.7%) 433(66.1%) 554(84.6%)
Meff.MWSL 4(0.6%) 14(2.2%) 40(6.2%) 74(11.4%) 126(19.4%) 195(30%) 282(43.4%) 416(64%) 416(64%)
Meff.Nyholt 34(5.2%) 121(18.6%) 239(36.8%) 331(50.9%) 419(64.5%) 491(75.5%) 543(83.5%) 588(90.5%) 638(98.2%)
Meff.Liji 18(2.8%) 64(9.8%) 133(20.5%) 195(30%) 263(40.5%) 320(49.2%) 365(56.2%) 413(63.5%) 382(58.8%)
Meff.Gao 65(10%) 131(20.2%) 213(32.8%) 275(42.3%) 327(50.3%) 372(57.2%) 412(63.4%) 449(69.1%) 317(48.8%)
Meff.Galwey 7(1.1%) 23(3.5%) 58(8.9%) 98(15.1%) 147(22.6%) 200(30.8%) 252(38.8%) 312(48%) 270(41.5%)
  1. R = ENT/ANT(%) ratio in brackets. ANT = 650. Meff estimates closest to ENT estimates in bold