Fig. 1
From: CANTARE: finding and visualizing network-based multi-omic predictive models
VOLARE-CANTARE workflow. a, b, and c are VOLARE steps. Using the resulting VOLARE network (c) as input, d and e are CANTARE steps. a Perform pairwise regressions across all analytes in all omes; n is the number of study participants in each ome, p is the number of predictors in each ome, and x is the number of pairwise regressions. b Visualize pairwise results, assessing credibility of underlying detail. Each circle represents one participant. The solid green line represents the fitted regression between the two analytes for controls. The dotted orange line represents the fitted regression for people with IBD. c Summarize the “top table” of promising results for each pair of omes in a single network. Each node represents an analyte; each edge represents one relationship from the top table. d Build predictive models from network neighborhoods. Shaded regions indicate models created from neighborhoods of order 2, centered on the indicated microbe. Nodes with black borders are included as predictors. e Visualize predictive models. The cumulative fit graph displays predicted probability by accumulating the contribution of one predictor at a time. Each path represents values for one study participant (green = control, orange = IBD)