Predicting Bioaccumulation of Nanomaterials: Modeling Approaches with Challenges

Understanding the bioaccumulation of nanomaterials (NMs) by organisms is essential in evaluating their potential ecotoxicity. However, the experimental determination of bioaccumulation is substantially challenging, which spawned the development of prediction approaches via establishing models for various NMs. Conventional modeling approaches, such as the biotic ligand model (BLM), partition coefficients, accumulation factor models, and quantitative structure-activity relationship (QSAR) models, were initially used in the application of NMs, aiming to provide a reliable quantitative dose in a resource-saving way. These methods, which are based on the uptake patterns of substances, probably lead to deviated results due to the different uptake behaviors of NMs. In this study, currently developed models to evaluate the bioaccumulation of NMs are critically reviewed, with their feasibilities and limitations being analyzed. In addition, the recently developed machine learning amended models have taken great efforts in realizing biological behaviors of NMs in organisms by providing <i>in silico</i> predictions. Though this data-driven approach has limitations in mechanism exploration, it may give different insights into the bioaccumulation model establishment and critical feature identification.