Fig.8
From: CANTARE: finding and visualizing network-based multi-omic predictive models
Distributions of predictors. Left panel: participant-level detail of predictors. Predictors across all models were normalized by range. Green squares represent controls, orange circles represent people with IBD. Jitter is used to reduce overplotting. Vertical lines indicate first and third quartile of predictor. Half of the observed values fall between these lines. Eubacterium siraeum and Oscilbacter splanchnicus are zero-heavy, with at least half of the observations being zero. Predictors flagged with an asterisk (*) are metabolic isomers. Parenthetical values after predictor names indicate which group of methods included the predictor, with c = CANTARE, p = penalized regression, r = random forest. After age and calprotectin, variables are listed alphabetically, within ome. Center and right panels: violin plots illustrating overall distribution of predictors, separated by IBD group