Fig. 4

CD location-specific protein pairs outperform calprotectin for inferring severity. A Log2-transformed Gelsolin protein abundance comparison between ICD (n = 25) and CCD (n = 19) abundance. ** denotes p < 0.01, Welch’s t test. B Scatterplot comparing CCD ratio (Gelsolin/Clostridia) pair value and SES-CD disease severity for ICD (ρ = 0.29, p = 0.21) and CCD (ρ = 0.77, p = 9e⁻⁵). Ratio input was log₂-scaled prior to ratio. Correlation values were obtained using Spearman’s correlation. SES-CD scores (X-axis) are scaled for visualization purposes. C Binned comparisons of CCD ratio pair abundance in ICD and CCD patients binned by “low” and “High” severity (median SES-CD split). ICD high sev. n = 8 and low sev. n = 17, CCD severity groups high patient n = 8, low patient n = 10. Statistical comparison performed using Welch’s t test, **p = 0.0033. D Log2-transformed Galectin-4 protein abundance comparison between ICD and CCD abundance. “NS” = non-significant difference (P > 0.05, Welch’s t test). E Scatterplot comparing ICD ratio pair value and SES-CD disease severity for ICD (ρ = 0.75, p = 1e⁻⁵⁾ and CCD (ρ = 0.14, p = 0.21). Ratio input was log₂-scaled prior to ratio. Correlation values were obtained using Spearman’s correlation. SES-CD scores (X-axis) are scaled for visualization purposes. F Binned comparisons of Galectin-4 abundance in ICD and CCD “low” and “high” severity median split groups. ICD severity groups split between high n = 8, low n = 17 and CCD severity groups split between high = 8, low = 10. Statistical comparison performed using Welch’s t test, ***p = 0.0002. G Log2-transformed feature abundance for all features used in ratios along with both Calprotectin subunits (S100A8 and 9). All samples (182) used to illustrate broad pattern of expression in comparison to calprotectin expression. H Table of statistics comparing the performance of location-specific ratios and calprotectin subunits. All statistics derived using Spearman correlational analysis