Fig. 4: Predictive ability and calibration of models including clinical, polygenic and/or metabolomic scores. | Nature Communications

Fig. 4: Predictive ability and calibration of models including clinical, polygenic and/or metabolomic scores.

From: Metabolomic and genomic prediction of common diseases in 700,217 participants in three national biobanks

Fig. 4

A Ten-year hazard ratios of models with clinical variables compared to those variables plus the best ‘omic data (either just metabolomics or metabolomics plus PGS from Fig. 2, n = 241,750). Dots represent Cox regression estimates and horizontal error bars denote 95% confidence intervals of the hazard ratio. B For each disease, the calibration of three-year observed event rates are shown by 10 equally sized deciles of absolute risk predicted by a metabolomic score adjusted for age and sex (n = 415,592). Dots represent means and vertical lines represent 95% confidence intervals of the observed event rate. Calibration slopes and intercepts were derived from a logistic regression of the observed risk on the predicted risk. Source data are provided as a Source Data file.

Back to article page