Fig. 5
Dose–response modelling and derivation of benchmark mark dose (BMD1SD) from the transcriptomics data. a Examples of dose–response modelling for several contigs showing the expression values (black dots) along the dose gradient. The dose–response model computed (red line) is used to determine the BMD1SD (vertical line), defined as the minimal dose for which the signal reach the benchmark response (BMR = y0 ± SD) with the modelled value at null dose and the SD the estimated residual standard deviation around the model (dotted horizontal lines). The ITDR (in μGy.h−1) as a log scale on the x-axis, responses at the control sites (G18 and H18) appear as half circles on the y-axis. b Distribution of the enriched GO biological process summarised by the 25th percentile of the BMD1SD values from all modelled contigs belonging to the considered GO term. The size of each point codes for the number of modelled contigs in each GO term. Asterisk symbol (*) indicates the 3 GO term for which empirical cumulative distribution function (ECDF) is shown as example in c. c Example of ECDF plots used to estimate BMD1SD shown in b (*). The 95% confidence intervals are indicated for each selected contigs categorised in 3 different GO biological processes. The dose (in μGy.h.−1) is indicated on the x-axis. The trend of the dose–response curves fitted for each contig is indicated by the colour code (bell: bell-shape, U: U-shape, dec: decreasing (downregulated), inc: increasing (upregulated)