Functionality and toxicity assessments of nanomaterials using machine learning

This thesis offers an interdisciplinary investigation that encompasses nanotechnology, microbiology, toxicology, and computer science, specifically, machine learning (ML), thereby highlighting the inherently collaborative nature of this research in the nanotechnology field. Understanding the unique properties of nanomaterials (NMs) is critical for predicting their potential toxicity and desired functionality while feeding and supporting the Safe and Sustainable by Design (SSbD) approach, which is currently gaining momentum. Predicting the functionality of NMs and their hazard potential involves identifying specific properties of NMs that are associated with toxicity and their functionality. For example, the size, shape, surface charge, and surface chemistry of NMs can all affect their toxicity on the one hand and their functionality on the other. We demonstrated that, all the assessments must take these properties into consideration, to enable better predictions. Through our research, it has become apparent that the present state of knowledge within the realm of nanotechnology is inadequate in facilitating an accurate assessment of the functionality/hazards of NMs, due to a number of challenges. There exist conspicuous scarcity of studies that have produced systematic and comprehensive datasets. However, in this work we advance the state-of-the art in the field of nanoinformatics by creating comprehensive datasets.