Threats to the soil microbiome from nanomaterials: A global meta and machine-learning analysis

Abstract

Soil is the primary sink for released nanomaterials (NMs), but the understanding of the impacts of NMs on the soil microbiome remains fragmented. Moreover, there is currently lack of systematic approaches to evaluate the microbial ecological risks of NMs. In this study, we conducted a global meta-analysis incorporating 2134 paired observations from 107 publications from 2000 to June 2023 to comprehensively assess the effects of NMs on the soil microbiome. Additionally, we developed a machine-learning approach to predict these impacts and identified key contributing features. The results reveal that NMs have significant negative effects on soil microbial diversity (−0.96%), biomass (−14.01%), activity (−3.39%), and function (−14.44%). The impacts of NMs on fungal diversity were greater than those on bacteria. Compared with carbon NMs, metal NMs have more pronounced negative effects on various soil microbial community metrics, with Ag NMs exhibiting the greatest negative impact. Ag NMs exhibited greater negative effects on microbial function than bulk Ag or Ag⁺. Nanoscale effects played a pivotal role in these adverse effects. These adverse effects are primarily associated with NM type, size and content. Two machine learning models achieved acceptable prediction accuracy in assessing the impact of NMs on the soil microbial community. This study offers an effective approach for the ecological risk assessment of NMs and provides a scientific foundation for the rational and informed application of NMs in the soil environment.