The Challenge
Modelling Nanomaterial Bioaccumulation
Understanding how engineered nanomaterials accumulate in organisms is critical for ecological risk assessment. Traditional biokinetic models struggle with the nonlinear uptake and elimination kinetics of nanomaterials. Parameter fitting for compartment models requires robust optimization approaches, and model selection between competing formulations demands rigorous statistical criteria.
Our Methodology
A rigorous multi-stage biokinetic modelling and optimization pipeline
One-Compartment Model (OC)
Implemented the classic one-compartment biokinetic model describing nanomaterial uptake and elimination with first-order kinetics. Computes uptake rate (ku), elimination rate (ke), and bioconcentration factor (BCF).
OC with Stored Fraction (OC-SF)
Extended with an OC-SF model accounting for a fraction of accumulated nanomaterial that is not readily eliminated — representing sequestration in tissues or organelles, important for persistent NMs like Ag and TiO2.
Genetic Algorithm Optimization
Implemented a genetic algorithm (using the Jenetics library in Java) for nonlinear regression parameter fitting. GA handles the complex, multimodal parameter spaces typical of nanomaterial biokinetic data better than traditional least-squares approaches.
Statistical Model Selection
Applied Akaike Information Criterion (AIC, AICc) and Akaike weights alongside adjusted R-squared to rigorously select between OC and OC-SF models. This ensures the most parsimonious model is chosen for each organism-NM combination.
Multi-Organism Validation
Validated across 8+ case studies including D. magna (Ag NMs, TiO2 NMs), earthworms (Ag, SiO2 NMs), and D. rerio (C60, graphene, GO, Au NMs), demonstrating broad applicability across organisms and nanomaterial types.