Figure 3: Re-scaling of MPO data for weighted-3D-SAMPgut and gene expression analyses in mice.

(a) MPO data from Fig. 2f is not normally distributed (P>0.1) and therefore not comparable for weighted analysis across SM-dissected tissues. Log-transformation allowed comparable normalization for ileum and colon (n=150 samples; B6, AKR, SAMP). (b) Re-scaling of MPO activity for adjustment of SM scores (percentage of abnormal mucosa). Log-transformation favours differentiation of samples with low MPO (steepest, below curve shoulder ∼200 U g⁻¹). (c) Raw qPCR-CT gene expression values correlation plots versus re-scaled MPO activity. In β-actin panel (narrowest variability), y=0.603x represents linear correlation slope (for each unit of increase in re-scaled-MPO, β-actin increased by 0.603 units); value 15.2 indicates intercept. Low R² indicates variability departing from fitted line cannot be explained by equation or MPO activity alone. (d) Example of SM-targeted 2^−ddCT-fold analysis of il6 and tlr4 genes (six samples per mouse) for two randomly selected dexamethasone-treated mice. Note reproducible regional expression profile variability within colon (tlr4:il6 ratio is highest distally, but lowest proximally), suggesting SM-guided studies are best by normal–abnormal paired sampling within the same region. Paired SM sampling of ileum shows also a pattern: high tlr4:il6 ratio in normal areas, but low in cobblestones. Paired sampling may reveal biologically relevant spatial/structural differences within the same mouse strain: for example, il6 is overexpressed in proximal colon, which is less inflamed in DSS-colitis, but also in small intestinal cobblestones which are markedly inflamed in SAMP ileitis. IL6 varies (anti/proinflammatory) in the same mouse strain depending on location and SM-lesion structure.