Performance evaluation in pairwise comparison tests using four methods (ANOVA, LIMMA, KW and Proposed) for the small-sample case.

<p>We generated 300 DE genes out of 20,000 total genes for <i>m</i> = 4 conditions with different patterns for a small-sample case (n1 = n2 = n3 = n4 = 6) and <i>σ</i>² = 0.05, with a 2-fold change in expression between the groups, to investigate the pattern-detection performance of the proposed method in comparison with the others. The values reported in the form {x, x, x, x} in this table represent the numbers of downregulated (DR) or upregulated (UR) differentially expressed (DE) genes estimated by the ANOVA, LIMMA, KW and proposed (Bold) methods, respectively. ^<i>a</i>Note that <math><mrow>log<msub><mi>μ</mi><mo>^</mo><mi>i</mi></msub><msub><mi>μ</mi><mo>^</mo><mi>j</mi></msub><mo><</mo><mo>−</mo><mn>1</mn></mrow></math> indicates significant 2-fold downregulation and <math><mrow>log<msub><mi>μ</mi><mo>^</mo><mi>i</mi></msub><msub><mi>μ</mi><mo>^</mo><mi>j</mi></msub><mo>></mo><mo>+</mo><mn>1</mn></mrow></math> indicates significant 2-fold upregulation.</p><p>Performance evaluation in pairwise comparison tests using four methods (ANOVA, LIMMA, KW and Proposed) for the small-sample case.