① Input
② Preview & Normalize
③ Pairwise
④ One-way ANOVA
⑤ LOGO
⑥ 2-way ANOVA
⑦ Results
Expected format
Row 1: sample names · Row 2: group labels · Rows 3+: numeric values. First column = feature ID. Tab-separated.
Item Sample1 Sample2 Sample3 Sample4
ID GroupA GroupA GroupB GroupB
Gene1 12.4 11.8 45.2 42.1
Gene2 8.3 9.1 8.7 9.2
Paste TSV data
📋
Paste your tab-separated data here
Ctrl+V / ⌘V
–
Features
–
Samples
–
Groups
Raw
Active Norm
Detected Groups
Normality Test (Shapiro-Wilk)
Run the test to see normality summary per group.
Interpretation: Use as diagnostic guidance only, not a binary decision gate. SW is underpowered (n<5) and over-sensitive (n>30). A high % passing supports parametric tests; low % suggests Welch's t-test (still robust) or Mann-Whitney U. Never apply t-test to features that "pass" and Mann-Whitney to those that "fail" — this creates selection bias.
Data Table (first 50 rows)
No data loaded yet.
Pairwise Comparison
Ratio = mean(A)/mean(B) · log₂FC = log₂(mean A + ε) − log₂(mean B + ε). Test type and correction set in sidebar.
VS
Options
One-Way ANOVA / Kruskal-Wallis
≥ 3 groups required
Data tested: –. Group means in results always shown in original (pre-test) scale.
Results appear in the Results tab
How to read ANOVA results: In Auto mode, each feature is independently tested for normality (Shapiro-Wilk) and variance homogeneity (Levene). Features passing both → ANOVA; passing normality but unequal variance → Welch's ANOVA; failing normality → Kruskal-Wallis. The Test column shows which test was applied. A significant omnibus result (adj.p < α) triggers post-hoc testing.
LOGO — Leave-One-Group-Out pairwise analysis
How LOGO works: For each group, all samples from the other groups are pooled into a single "rest" group. Welch's t-test compares each group vs. its rest. Significant features indicate that a group differs from the global average. FDR correction is applied across all features × comparisons.
Run LOGO analysis first.
Two-Way ANOVA
Load data first, then assign Factor B levels below.
How to read 2-way ANOVA: Reports F-statistics and adj. p-values for Factor A (group effect), Factor B (second factor effect), and A×B interaction. A significant interaction means the effect of Factor A depends on the level of Factor B. Post-hoc tests are run for Factor A within each level of Factor B.
Configure Factor B and run analysis.
No results yet.
Results Table
Run a pairwise comparison first.