Extended Data Fig. 3: Related to Fig. 1. (See Supplementary Table 12 for statistics).

(a - d) Plots showing the effect of fructose intake (kJ/mouse/day) on insulin tolerance (AUC) (a) of mice at 15-16 weeks and fasting blood glucose concentration (mmol/l) (b), fasting blood insulin concentration (ng/ml) (c) and their product (mmol/l x ng/ml) (d) at 14 weeks. The relationship between metabolic parameters and fructose intake is shown at low (10% energy; 1.43 kJ/g), medium (20%; 2.86 kJ/g) and high (30%; 4.29 kJ/g) dietary protein content. As the fructose intake increases along the x-axis, the amount of glucose eaten decreases. Each symbol (o) represents an individual mouse (n = 292, 293, 288 and 288 mice for a, b, c and d respectively). The fitted lines are derived from data analysed by GAM, fitting an interaction between a smooth term for dietary fructose intake (in one carbohydrate dimension) and protein content as a three-level categorical factor and the dotted lines represent s.e.m. for fitted values. (e) Response surfaces showing the relationship between the intake of fructose, glucose and protein derived energy (kJ/mouse/day) and product of fasting blood glucose and fasting blood insulin concentration (mmol/l x ng/ml) at 6 weeks. (f) Relationship between fat mass (grams) and insulin sensitivity (fasting blood glucose (mmol/l) x fasting insulin (ng/ml)) measured at 13-14 weeks (n = 288 mice). R² and P value (P = 2.37E-31) for linear regression of data are shown. (g) Plot showing the effect of fructose intake (kJ/mouse/day) on glucose tolerance (AUC) of mice at 14 weeks. The relationship between glucose tolerance and fructose intake is shown at low (10%), medium (20%) and high (30%) dietary protein content. Each symbol (o) represents an individual mouse (n = 293 mice). The fitted lines are derived from GAM and the dashed lines represent s.e.m. for fitted values. (h, i) Response surfaces showing the relationship between the intake of fructose, glucose and protein derived energy (kJ/mouse/day) and total AUC calculated from the glucose tolerance test (h) at 6 weeks and peak blood insulin concentrations (ng/ml) (i) (15 minutes after glucose administration) measured at 14 weeks.