Fig. 1

The metabolomic model of RSV bronchiolitis versus healthy control children. Urine metabolite levels were measured in age matched healthy children and compared to those with RSV bronchiolitis in the ED. Using a blind test set, PLS-DA analysis (SIMCA P-11) of these metabolites created a model of separation (R2 = 0.86, Q2 = 0.76). Illustrated are: a the Variables of Importance plot ranking the metabolites according to their significance in the model; b scaled and centered metabolite differences between groups shown as the Coefficients of Variation plot; c the PLS-DA prediction scores for each subject with error bars representing means and 95 % confidence intervals. The PLS-DA algorithm separates groups of data based on a score of 0–1; in this case a value above 0.5 indicates the infant is healthy while below 0.5 indicates RSV bronchiolitis