Fig. 3: Limma and ComBat-based HarmonizR application for batch effect reduction across different experimental time points and tissue preservations.
a Scheme of the experimental design. b Batch count distribution of all 3530 proteins quantified at least 2 times in a batch. c Heatmap visualization of Pearson correlation-based hierarchical clustering with Ward.D linkage for each tissue type and timepoint for unharmonized combined data, after ComBat- and limma- based HarmonizR execution. Sample specific CV and mean are shown on lower panels. (Batch 1 (green): n = 6 biologically independent animals (Tumor: n = 3; Control: n = 3); Batch 2 (blue): n = 5 biologically independent animals (Tumor: n = 2; Control: n = 3); Batch 3 (pink): n = 7 biologically independent animals (Tumor: n = 5; Control: n = 2); Batch 4 (turquoise): n = 7 biologically independent animals (Tumor: n = 5; Control: n = 2)). In boxplots, 50% of the data points are inside the box (Q1 (Quartile 1) being the lower bound of the box (25%), Q3 being the upper bound of the box (75%)). Whiskers show all values beyond the box without outliers. Outliners were defined as Q3 + 1.5 * IQR (Interquartile range) (upper outlier) and Q1-1.5 * IQR (lower outlier). IQR being Q1–Q3. d Batch specific coverage of proteins, associated with the “REACTOME-Signaling by Hedgehog” gene set. e Abundance distribution of the Sonic Hedgehog medulloblastoma Marker Filamin A in unnormalized data and after ComBat and limma- based HarmonizR execution. f Overlap of p-value significant proteins (p-value < 0.05), identified in t-testing between cerebellar tumors of hGFAP-cre::SmoM2Fl/+ mice and control cerebella in unnormalized data and after ComBat- and limma- based HarmonizR execution (Two-sample Student’s T-test, p-value < 0.05). Source data are provided as a Source Data file.
