FIGURE LEGENDS
Figure 1. Assessment of renal injury in untreated (black), saline-treated (blue), and cisplatin-treated (red) C57BL/6 mice (A, B ), and FVB/N mice (C, D ). Plasma creatinine concentrations (A, C ) were normalized to creatinine-d3 standard before quantification. (B, D ) H&E-stained kidney sections were scored from 0-5 based on severity of renal injury by a pathologist blinded to treatment conditions. n=5-9 (exact sample sizes can be found in Table S1 ), error bars represent mean ± SD. Statistical analysis was performed using two-way ANOVA, *p < 0.05 compared to saline control.
Figure 2. Orthogonal partial least squares discriminant analysis scores plots comparing plasma (A-D ), urine (E-H ), and kidney (I-L ) samples from saline-treated (black) and cisplatin-treated (red) C57BL/6 mice at each timepoint: day 1 (A, E, I ), day 2 (B, F, J ), day 3 (C, G, K ), day 4 (D, H, L ). Model statistics for OPLS-DA were as follows: (A ) R2Y = 0.989, Q2Y = 0.369 (B ) R2Y = 0.980, Q2Y = 0.542 (C ) R2Y = 0.989, Q2Y = 0.467 (D ) R2Y = 0.985, Q2Y = 0.580 (E ) R2Y = 0.627, Q2Y = 0.033 (F ) R2Y = 0.989, Q2Y = 0.649 (G ) R2Y = 0.989, Q2Y = 0.295 (H ) R2Y = 0.946, Q2Y = 0.581 (I ) R2Y = 0.914, Q2Y = 0.101 (J ) R2Y = 0.979, Q2Y = 0.611 (K ) R2Y = 0.944, Q2Y = 0.426 (L ) R2Y = 0.973, Q2Y = 0.569. n=5-10, exact sample sizes can be found in Table S1 .
Figure 3. Time course cluster analysis of (A) plasma, (B) urine, and (C) kidney features that were significantly altered by cisplatin treatment in C57BL/6 mice. The lighter, thinner lines represent change over time for each individual feature within each cluster, and the darker, thicker line represents the mean z-score for all features in each cluster. Boxes represent the median and interquartile range (IQR), and the whiskers represent 1.5x IQR.
Figure 4. Log transformed relative intensity of features significantly affected by cisplatin treatment in plasma samples from C57BL/6 mice. Data are presented as box and whisker plots, where boxes represent the median and interquartile range (IQR), and whiskers represent 1.5x IQR. Statistical analysis was performed using two-way ANOVA, *p<0.05 compared to saline control at each timepoint. n=5-9, exact sample sizes can be found in Table S1 .
Figure 5. Log transformed relative intensity of features significantly affected by cisplatin treatment in urine samples from C57BL/6 mice. Data are presented as box and whisker plots, where boxes represent the median and interquartile range (IQR), and whiskers represent 1.5x IQR. Statistical analysis was performed using two-way ANOVA, *p<0.05 compared to saline control at each timepoint. n=6-9, exact sample sizes can be found in Table S1 .
Figure 6. Log transformed relative intensity of features significantly affected by cisplatin treatment in kidney samples from C57BL/6 mice. Data are presented as box and whisker plots, where boxes represent the median and interquartile range (IQR), and whiskers represent 1.5x IQR. Statistical analysis was performed using two-way ANOVA, *p<0.05 compared to saline control at each timepoint. n=7-10, exact sample sizes can be found in Table S1 .
Figure 7. Receiver operating characteristic (ROC) analysis of early plasma biomarkers. (A ) Summary of area under the ROC curve values calculated for each plasma metabolite at all timepoints. (B, C ) Multivariate ROC curves generated at (B ) day 1 and (C ) day 2 using a variable number of metabolites. ROC curves were generated using a sequential combination of the top two to top ten ranked metabolites, where each metabolite was ranked using the calculated univariate AUROC at each respective timepoint.
Figure 8. Receiver operating characteristic (ROC) analysis of early urine biomarkers in C57BL/6 mice. (A ) Summary of area under the ROC curve values calculated for each urine metabolite at all timepoints. (B, C ) Multivariate ROC curves generated at (B ) day 1 and (C ) day 2 using a variable number of metabolites. ROC curves were generated using a sequential combination of the top two to top 15 ranked metabolites, where each metabolite was ranked using the calculated univariate AUROC at each respective timepoint.