Biology subjects

CHARLES, S.

Publications and source records attributed to CHARLES, S..

7 recordsLinked to original sources

bioRxiv2025

A Bayesian Neural Ordinary Differential Equations Framework to Study the Effects of Chemical Mixtures on Survival

In the field of Plant Protection Products (PPP), the combination of multiple active ingredients is a common strategy to improve the efficacy against target species. However, these mixtures may expose non-target species to untested chemical combinations due to varied application and degradation pathways. Developing a method to predict the effects of such mixtures without extensive testing is crucial for enhancing risk assessment for non-target organisms. Toxicokinetic (TK) and toxicodynamic (TD) models currently provide predictive power for the toxicity of mixture over time, utilizing the concentration addition and independent action principles. However, their capacity to capture complex, non-linear interactions -- such as synergistic and antagonistic effects -- remains limited. To overcome these limitations, we introduce a new methodology that integrates TKTD models using Ordinary Differential Equations (ODE) with neural networks (NN), offering a robust framework for modeling complex substance interactions. The ODE component encodes fundamental biological principles that constrain the neural network to biologically plausible solutions. Bayesian inference further refines the model by addressing uncertainties in data, model parameters, and biological processes, while also allowing quantification of prediction uncertainty, highlighting gaps in experimental data that warrant further investigation. This hybrid model was evaluated across 99 acute toxicity studies that included various PPP mixtures, testing its ability to identify and forecast deviations from expected mixture behaviors. Our findings highlight the potential of combining mechanistic models with machine learning techniques to advance predictive accuracy in environmental toxicology, providing a framework to anticipate the combined effects of untested chemical mixtures and inform risk assessment practices.

pharmacology and toxicology↗

bioRxiv2022

Chemical effects on ecological interactions within a model-experiment loop

We propose in this paper a method to assess the effects of a contaminant on a micro-ecosystem, integrating the population dynamics and the interactions between species. For that, we developed a dynamic model to describe the functioning of a microcosm exposed to a contaminant and to discriminate direct and indirect effects. Then, we get back from modelling to experimentation in order to identify which of the collected data have really been necessary and sufficient to estimate model parameters in order to propose a more efficient experimental design for further investigations. We illustrated our approach using a 2-L laboratory microcosm involving three species (the duckweed Lemna minor, the microalgae Pseudokirchneriella subcapitata and the daphnids Daphnia magna) exposed to cadmium contamination. We modelled the dynamics of the three species and their interactions using a mechanistic model based on coupled ordinary differential equations. The main processes occurring in this three-species microcosm were thus formalized, including growth and settling of algae, growth of duckweeds, interspecific competition between algae and duckweeds, growth, survival and grazing of daphnids, as well as cadmium effects. We estimated model parameters by Bayesian inference, using simultaneously all the data issued from multiple laboratory experiments specifically conducted for this study. Cadmium concentrations ranged between 0 and 50 g.L-1 . For all parameters of our model, we obtained biologically realistic values and reasonable uncertainties. The cascade of cadmium effects, both direct and indirect, was identified. Critical effect concentrations were provided for the life history traits of each species. An example of experimental design adapted to this kind a microcosm was also proposed. This approach appears promising when studying contaminant effects on ecosystem functioning.

ecology↗

bioRxiv2021

rbioacc: an R-package to analyse toxicokinetic data

O_LI rbioacc is an R-package dedicated to the analysis of experimental data collected from bioaccumulation tests during which organisms are exposed to a chemical (exposure phase) and then put into a clean media (depuration phase). Internal concentrations are regularly measured over time all along the experiment. C_LIO_LI rbioacc provides ready-to-use functions to visualize and fully analyze such data. Under a Bayesian framework, this package fits a generic one-compartment toxicokinetic (TK) model automatically built from the data. It provides TK parameter estimates (appropriate uptake and elimination rates) and bioaccumulation metrics (e.g., BCF, BSAF, BMF). All parameter estimates, bioaccumulation metrics as well as predictions of internal concentrations into organisms are delivered with their uncertainty. C_LIO_LIBioaccumulation metrics are primarily provided in support of environmental risk assessment, in full compliance with regulatory requirements required to approve marketing applications of chemical substances. C_LIO_LIThis paper gives brief worked examples of the use of rbioacc from data collected through standard bioaccumulation tests, and publicly available within the scientific literature. These examples constitute step-by-step user-guides to analyze any new data set, uploaded in the right format. C_LI

ecology↗

bioRxiv2021

Generic solving of one-compartment toxicokinetic models

This paper gives the full analytical solution of the generic set of ordinary differential equations that define one-compartment toxicokinetic models. These models describe uptake and elimination processes taking place within living organisms when exposed to chemical substances. The models solved in this paper consider living organisms as a unique compartment, into which a parent compound enters via several possible exposure routes and from which it is eliminated as well as its potential metabolites. Benefiting from generic solutions of one-compartment toxicokinetic models is particularly useful when fitting them to experimental data, facilitating the writing of the inference algorithms leading to parameter estimates. Additionally, these models are of crucial interest in environmental risk assessment for the calculation of bioaccumulation metrics as required by regulators in support of decision making when they evaluate dossiers for marketing authorisation of active substances. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=129 SRC="FIGDIR/small/442956v2_ufig1.gif" ALT="Figure 1"> View larger version (18K): org.highwire.dtl.DTLVardef@1e8381org.highwire.dtl.DTLVardef@14b957forg.highwire.dtl.DTLVardef@143dc14org.highwire.dtl.DTLVardef@25c478_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphical abstractC_FLOATNO C_FIG

pharmacology and toxicology↗

bioRxiv2021

Accumulation-depuration data collection in support of toxicokinetic modelling

Regulatory bodies requires evaluation of bioaccumulation of chemicals within organisms with the objective to better assess risks linked to the toxicity of active substances. To this end, toxicokinetic (TK) data are particularly useful to relate the chemical exposure concentration to the accumulation and depuration processes happening within organisms. The bioaccumulative property of substances is quantified by bioaccumulation metrics obtained by fitting TK models to data collected from bioaccumulation tests. The internal concentrations of the studied substances are measured within organisms at regular time points during both accumulation and depuration phases, and their time course is captured by TK models thus providing bioaccumulation metrics. Still today, raw TK data remain difficult to access, as most of the time provided within papers in plots only. To increase accessibility to TK data, we present in this paper a wide collection of raw data sets extracted from the scientific literature in support of TK modelling to be performed with the MOSAICbioacc web application (https://mosaic.univ-lyon1.fr/bioacc/).

ecology↗

bioRxiv2021

TKTDsimulation.jl and tktdjl2r: innovative packages for High Performance Computing of survival predictions in support of environmental risk assessment under time-variable scenarios

Predictive environmental risk scenarios are today of major interest for environmental risk assessment as they provide plausible and consistent descriptions of possible effects of chemical in natura. In particular, they can be used for predictions of the future as consistent descriptions of pathways towards desired targets to protect. One single scenario would therefore be meaningless, as it could not capture all the variability and uncertainty involved in natural phenomenon combined with socio-economical events. A set of environmental risk scenarios is then a key asset to address sustainable and collaborative decision making associated with appropriate actions. Toxicokinetics-Toxicodynamics (TKTD) models are increasingly used for the assessment and the prediction of environmental risk assessment due to chemical products. This mechanistic modelling approach offers many advantages as the possibility to perform simulations under non-observed realistic situations with time-variable exposure profiles embedded in environmental risk scenarios. TKTD simulations can also be linked with other types of models (e.g., Individual Based Model) within a pipeline of computing inference as for example Bayesian inference or Machine Learning. To handle such challenges within the particular framework of TKTD models for survival, we present an innovative simulation tool written in the new programming language Julia, called TKTDsimulations.jl. Given that TKTD models for survival usually require high performance computing due to the numerical integration of differential equations, our tool strongly benefits from Julias facilities, in particular a code that is fast to compile and easy to maintain. In addition, to ease the link with the already developed R-package morse dedicated to the statistical handling of ecotoxicity data, we also developed a new R-package, called tktdjl2r, interfacing morse with our new simulation tool TKTDsimulations.jl that considerably faster predictions with the corresponding ready-to-use morse functions.

ecology↗

bioRxiv2020

How to account for the uncertainty from standard toxicity tests in species sensitivity distributions: an example in non-target plants

This research aims to account for the uncertainty on 50% effective rates (ER50) in species sensitivity distribution (SSD) analyses and to study how including this uncertainty may influence the 5% Hazard Rate (HR5) estimation. We explored various endpoints (survival, emergence, shoot dry weight) for non-target plants from seven standard greenhouse studies that used different experimental approaches (vegetative vigour vs. seedling emergence) and applied seven herbicides at different growth stages. Firstly for each endpoint of each study, a three-parameter log-logistic model was fitted to experimental toxicity test data for each species under a Bayesian framework to get a posterior probability distribution for ER50. Then in order to account for the uncertainty on the ER50, we explored two censoring criteria to censor ER50 taking the ER50 distribution and the range of tested rates into account. Based on dose-response fitting results and censoring criteria, we considered input ER50 values SSD analyses in three ways (only point estimates chosen as ER50 medians, interval-censored ER50 based on their 95% credible interval and censored ER50 according to one of the two criteria), by fitting a log-normal distribution under a frequentist framework to get the three corresponding HR5 estimates. We observed that SSD fitted reasonably well when there were at least six distinct ER50 values. By comparing the three SSD curves and the three HR5 estimates, we found that propagating the uncertainty from ER50 and including censored data into the SSD analysis often leads to smaller point estimates of HR5, which is more conservative in a risk assessment context. In addition, we recommend not to focus solely on the point estimate of the HR5, but also to look at the precision of this estimate as depicted by the 95% confidence interval.Competing Interest StatementThe authors have declared no competing interest.View Full Text

ecology↗