Skip to main content

Table 1 Comparison of screening and clustering results (low noise)

From: A method to identify differential expression profiles of time-course gene data with Fourier transformation

    Without screening With screening
AR(1) parameter Method J Error Sil ARI Error Sil ARI Sensitivity Specificity FDR FNR
p = 0.1 FC* 2 .037 .509 .909 .015 .560 .918 .878 .723 .121 .276
  3 .020 .471 .921 .016 .484 .932 .860 .783 .139 .216
  4 .015 .430 .963 .017 .438 .937 .863 .842 .136 .157
  5 .015 .388 .964 .014 .403 .944 .854 .798 .145 .201
  8 .015 .305 .964 .017 .317 .940 .851 .836 .148 .163
GPR**     .855 .779 .220 .144
p = 0.2 FC 2 .052 .471 .871 .021 .523 .875 .871 .722 .128 .277
  3 .036 .423 .912 .026 .443 .888 .846 .783 .153 .217
  4 .029 .386 .931 .028 .398 .895 .847 .839 .152 .160
  5 .027 .348 .935 .022 .366 .906 .837 .798 .162 .205
  8 .028 .274 .936 .029 .287 .895 .830 .836 .169 .163
GPR     .826 678 .321 .173
p = 0.3 FC 2 .073 .430 .822 .030 .487 .815 .863 .723 .136 .276
  3 .056 .380 .865 .042 .402 .814 .828 .783 .171 .217
  4 .052 .339 .875 .045 .356 .825 .827 .834 .172 .165
  5 .049 .306 .883 .036 .326 .845 .817 .790 .182 .209
  8 .047 .244 .888 .049 .257 .823 .803 .832 .196 .167
GPR     .798 .571 .428 .201
p = 0.5 FC 2 .159 .340 .610 .056 .414 .633 .835 .717 .165 .201
  3 .139 .287 .663 .093 .329 .591 .775 .768 .224 .231
  4 .124 .255 .702 .113 .280 .578 .766 .811 .233 .188
  5 .132 .226 .682 .093 .259 .615 .762 .773 .237 .226
  8 .143 .181 .649 .134 .205 .562 .730 .815 .269 .184
GPR     .756 .410 .589 .244
p = 0.7 FC 2 .266 .287 .345 .088 .357 .351 .755 .704 .244 .295
  3 .264 .224 .347 .153 .272 .314 .682 .738 .317 .261
  4 .258 .190 .370 .186 .230 .303 .668 .771 .331 .228
  5 .258 .171 .375 .161 .211 .317 .676 .745 .324 .255
  8 .267 .137 .339 .220 .172 .287 .641 .769 .358 .230
  GPR     .731 .335 .664 .268
  1. * FC: proposed method with Fourier coefficients, **GPR: Gaussian process regression.
  2. Comparison of estimation error rate (E), Silhouette width (S) and Adjusted Rand Index (ARI) values of model-based clustering without screening vs with screening with J Fourier coefficients including sensitivity, specificity, FDR and FNR with m = 20 time points. These summaries are based on 500 repetitions of each consisting of 800 curves with AR(1) parameter ρ’s with the noise standard deviation σ = 0.5.