Table 1.
Applicability of bootstrap methods to assays with only one measurement per concentration, determination of a winning model, and calculating a hit call probability.
Fig 1.
Comparison of normal distribution with standard deviation equal to bmad (orange line) and the empirical cumulative distribution function (ecdf) for points used to calculated bmad (black line).
For each assay, the bmad is calculated as the scaled mad of the response values for the lowest two concentrations per chemical. Deviations between the ecdf and the normal distribution at higher response values can be attributed to highly potent chemicals with a biological response at the lowest two concentrations as well as sources of noise that are from a non-normally distributed process.
Fig 2.
Analysis of bootstrap distribution of model parameters for bisphenol AF in ATG_ERa_TRANS_up assay.
A-C: Values for the hill model log(AC50) (A), coefficient (B), and top (C) parameters for all 1000 bootstrap samples. D-F: Values for the gain loss model gain log(AC50) (D), gain coefficient (E), and top (F) for all 1000 bootstrap samples. G-I: Values for the winning model gain log(AC50) (G), gain coefficient (H), and top (I) for all 1000 bootstrap samples, colored by winning model (hill = red, gain loss = blue). J: Correlation plot of winning model top vs. winning model gain log(AC50), colored by winning model (hill = red, gain loss = blue). K: Normalized experimentally measured values (black circles) and winning model (gain loss, black curve). Subset of fitted bootstrap resamples, with winning hill (red lines) and gain loss (blue lines) models plotted. Horizontal black lines represent 3x bmad (dashed) and activity cutoff (solid). L: Comparison to results from other assays. Cumulative empirical distribution function of winning model gain log(AC50) value for all bisphenol AF samples in all assays where the experiment results were determined to be a positive hit. Curves are colored by assay source, with TOX21 black, NVS orange, ATG sky blue, OT bluish green, and ACEA yellow.
Fig 3.
Model selection in hit call probability for sixteen estrogen receptor agonist assays.
For each plot, chemicals are ordered on the x-axis based on their hit call probability. The y axis indicates the percent of bootstrap resamples that were calculated to be a positive hit with a hill model (red), gain loss (blue), or a negative hit (black).
Fig 4.
Estrogen receptor model AUC values for chemicals with an AUC(Agonist) value > 0.1.
Point estimates for agonist (red), antagonist (black), and pseudoreceptor (blue) values are marked by circles for all AUC values with an upper 95% confidence interval > 0.1. Error bars indicate the 95% confidence interval obtained by bootstrap resampling.
Fig 5.
Nordihydroguaiaretic acid bootstrap curves.
Each of the 18 ER assays are shown in a separate panel with the assay cutoff indicated with a dashed horizontal line. Circles represent the pipeline normalized concentration-response data and the solid black line indicates the winning model fit to the data if the hit call was positive. TOX21_ERa_LUC_BG1_Antagonist was not a hit in the pipeline therefore no black line is drawn. All bootstrap curves with a positive hit call are drawn with hill and gns models colored red and blue respectively. All assays had a 100% hit call in the bootstrap results except for ACEA_T47D_80hr_Positive where 602 of the 1000 samples had a positive hit call and assays ATG_ERa_TRANS_up and TOX21_ERa_LUC_BG1_Agonist where a single bootstrap replicate in each assay was inactive.
Fig 6.
Estrogen receptor model agonist AUC values for chemicals tested at least twice in the uterotrophic assay.
Point estimates for agonist are colored by the uterotrophic consensus result being positive (red), equivocal (blue), and negative (black). Equivocal results in the uterotrophic assay indicate some tests were positive while others negative. Error bars indicate the 95% confidence interval obtained by bootstrap resampling.
Table 2.
Estrogen receptor assays included in this study.
Fig 7.
Normalized experimental concentration-response points (cyan circles) and corresponding hill model (cyan line) are shown. The distribution of smooth bootstrap resampled points (black circles) and fitted values (black lines) are indicated, highlighting the range of resampled observations for response values and the subsequent possibilities for the fitted hill model.