Background Minimal erythema dose (MED) has substantial inter‐ and intraindividual variations, reflecting the influence of very diverse factors. However, related studies showed little consistency probably because of their limited sample size. Objective To identify the factors associated with MED variations in a large‐scale population study. Methods The MED test was performed by following the international standard procedure on 22 146 subjects. The results were analysed in adjusted multivariable linear and logistic regression models. Results This large‐scale study revealed that lower MED was consistently associated with lighter skin [β‐coefficient = −0.33, 95% confidence interval ( CI ) −0.36 to 0.30, P = 6.41 × 10 −84 ]. Females had significantly higher MED than male (β = 0.91, 0.32–1.50, P = 2.93 × 10 −3 ). Stratified analyses showed that MED was not associated with age [female: odds ratio ( OR ) = 0.99, 0.98–1.01; male: OR = 0.99, 0.97–1.00]. MED was lower in summer than in other seasons (spring: OR = 1.08, 1.06–1.11; autumn: OR = 1.11, 1.08–1.13; winter: OR = 1.20, 1.18–1.22). Furthermore, MED was associated with air temperature (β = −0.36, −0.49 to 0.23, P = 4.81 × 10 −8 ) and air pressure (β = −0.64, −0.82 to 0.46, P = 8.01 × 10 −12 ) in summer only while not in other seasons. Conclusions This study provides unprecedented evidence that MED is associated with skin colour, sex, season and meteorological factors, but not with age.

The potential toxicity of chemicals may present adverse effects to the environment and human health. The quantitative structure-activity relationship (QSAR) provides a useful method for hazard assessment. In this study, we constructed a QSAR model based on a highly heterogeneous data set of 571 compounds from the US Environmental Protection Agency, for predicting acute toxicity to the fathead minnow (Pimephales promelas). An approach coupling support vector regression (SVR) with the genetic algorithm (GA) was developed to build the model. The generated QSAR model showed excellent data fitting and prediction abilities: the squared correlation coefficients (r(2)) for the training set and the test set were 0.826 and 0.802, respectively. Only eight critical descriptors, most of which are closely related to the toxicity mechanism, were chosen by GA-SVR, making the derived model readily interpretable. In summary, the successful case reported here highlights that our GA-SVR approach can be used as a general machine learning method for toxicity prediction.

No abstract