https://orcid.org/0000-0001-6990-9851
-
Views
-
Cite
Cite
Renske P.J. Hoondert, Rik Oldenkamp, Dick de Zwart, Dik van de Meent, Leo Posthuma, Reply to “Concerns About Reproducibility, Use of the Akaike Information Criterion, and Related Issues in Hoondert et al. 2019” and Focus in Developing QSAR‐Based Species Sensitivity Distributions, Environmental Toxicology and Chemistry, Volume 39, Issue 7, 1 July 2020, Pages 1302–1304, https://doi.org/10.1002/etc.4737
Close - Share Icon Share
Extract
Authors' Response:
Our exploratory study on the quantitative structure–activity relationship (QSAR)‐based species sensitivity distributions (SSDs)—which boils down to directly deriving SSDs for untested compounds from those of tested compounds (Hoondert et al. 2019)—has been appreciated by Iwasaki and Hayashi (2020). However, they also voice concerns about the statistics used in the study, particularly concerning the improper use of statistical methods and parameters in selecting appropriate models and the reproducibility of outcomes. Herewith we provide our reply to the comments, hoping to clarify our methods and reemphasize the higher aim of the study, which is to help forward this field to maturity.
ON NEEDS AND WIDER
There is a societal need to handle >350 000 chemicals and their mixtures, protecting environmental quality and human health when possible and restoring them when needed (Wang et al. 2020) In their comments, Iwasaki and Hayashi first mention the derivation of protective standards as a key goal of ecological risk assessment, hinting at the need for QSARs to bridge data gaps in the light of limited ecotoxicity data. To be clear, we argue that the utility of the methods we proposed is wider, with ecotoxicity data serving in established operational methods for environmental protection, life cycle impact assessment, and environmental pollution assessment and management (Hoondert et al. 2019; Posthuma et al. 2019) We also foresee opportunities to support the development of sustainable chemistry methods (“benign by design” chemicals; Blum et al. 2017) Furthermore, looking at, for example, the US Environmental Protection Agency's (2020) ecotoxicity database alone, these methods can be based on nearly a million ecotoxicity endpoint data, covering tests for >12 000 chemicals, >13 000 species, representing >50 000 references. Wider policy needs and practices can be based on many data.
