The rapid increase in both the quantity and complexity of data that are being generated daily in the field of environmental science and engineering (ESE) demands accompanied advancement in data analytics. Advanced data analysis approaches, such as machine learning (ML), have become indispensable tools for revealing hidden patterns or deducing correlations for which conventional analytical methods face limitations or challenges. However, ML concepts and practices have not been widely utilized by researchers in ESE. This feature explores the potential of ML to revolutionize data analysis and modeling in the ESE field, and covers the essential knowledge needed for such applications. First, we use five examples to illustrate how ML addresses complex ESE problems. We then summarize four major types of applications of ML in ESE: making predictions; extracting feature importance; detecting anomalies; and discovering new materials or chemicals. Next, we introduce the essential knowledge required and current shortcomings in ML applications in ESE, with a focus on three important but often overlooked components when applying ML: correct model development, proper model interpretation, and sound applicability analysis. Finally, we discuss challenges and future opportunities in the application of ML tools in ESE to highlight the potential of ML in this field.

The greater abundances of antibiotic resistance genes (ARGs) in point-of-use tap and reclaimed water than that in freshly treated water raise the question whether residual disinfectants in distribution systems facilitate the spread of ARGs. This study investigated three widely used disinfectants (free chlorine, chloramine, and hydrogen peroxide) on promoting ARGs transfer within Escherichia coli strains and across genera from Escherichia coli to Salmonella typhimurium. The results demonstrated that subinhibitory concentrations (lower than minimum inhibitory concentrations [MICs]) of these disinfectants, namely 0.1-1 mg/L Cl for free chlorine, 0.1-1 mg/L Cl for chloramine, and 0.24-3 mg/L HO, led to concentration-dependent increases in intragenera conjugative transfer by 3.4-6.4, 1.9-7.5, and 1.4-5.4 folds compared with controls, respectively. By comparison, the intergenera conjugative frequencies were slightly increased by approximately 1.4-2.3 folds compared with controls. However, exposure to disinfectants concentrations higher than MICs significantly suppressed conjugative transfer. This study provided evidence and insights into possible underlying mechanisms for enhanced conjugative transfer, which involved intracellular reactive oxygen species formation, SOS response, increased cell membrane permeability, and altered expressions of conjugation-relevant genes. The results suggest that certain oxidative chemicals, such as disinfectants, accelerate ARGs transfer and therefore justify motivations in evaluating disinfection alternatives for controlling antibiotic resistance. This study also triggers questions regarding the potential role of environmental chemicals in the global spread of antibiotic resistance.

Although widespread antibiotic resistance has been mostly attributed to the selective pressure generated by overuse and misuse of antibiotics, recent growing evidence suggests that chemicals other than antibiotics, such as certain metals, can also select and stimulate antibiotic resistance via both co-resistance and cross-resistance mechanisms. For instance, tetL, merE, and oprD genes are resistant to both antibiotics and metals. However, the potential de novo resistance induced by heavy metals at environmentally-relevant low concentrations (much below theminimum inhibitory concentrations [MICs], also referred as sub-inhibitory) has hardly been explored. This study investigated and revealed that heavy metals, namely Cu(II), Ag(I), Cr(VI), and Zn(II), at environmentally-relevant and sub-inhibitory concentrations, promoted conjugative transfer of antibiotic resistance genes (ARGs) between E. coli strains. The mechanisms of this phenomenon were further explored, which involved intracellular reactive oxygen species (ROS) formation, SOS response, increased cell membrane permeability, and altered expression of conjugation-relevant genes. These findings suggest that sub-inhibitory levels of heavy metals that widely present in various environments contribute to the resistance phenomena via facilitating horizontal transfer of ARGs. This study provides evidence from multiple aspects implicating the ecological effect of low levels of heavy metals on antibiotic resistance dissemination and highlights the urgency of strengthening efficacious policy and technology to control metal pollutants in the environments.

This study performed mechanistic toxicity assessment of nanosilver (nAg) and nanotitanium dioxide anatase (nTiO2_a) via toxicogenomic approach, employing a whole-cell-array library consisting of 91 recombinated Escherichia coli K12 strains with transcriptional GFP-fusions covering most known stress response genes. The results, for the first time, revealed more detailed transcriptional information on the toxic mechanism of nAg and nTiO2_a, and led to a better understanding of the mode of action (MOA) of metal and metal oxide nanomaterials (NMs). The detailed pathways network established for the oxidative stress system and for the SOS (DNA damage) repair system based on the temporal gene expression profiling data revealed the relationships and sequences of key genes involved in these toxin response systems. Both NMs were found to cause oxidative stress as well as cell membrane and transportation damage. Genotoxicity and DNA damage were also observed, although nTiO2_a induced SOS response via previously identified pathway and nAg seemed to induce DNA repair via a pathway different from SOS. We observed that the NMs at lower concentration tend to induce more chemical-specific toxicity response, while at higher concentrations, more general global stress response dominates. The information-rich real-time gene expression data allowed for identification of potential biomarkers that can be employed for specific toxin detection and biosensor developments. The concentration-dependent gene expression response led to the determination of the No Observed Transcriptional Effect Level (NOTEL) values, which can be potentially applied in the regulatory and risk assessment framework as an alternative toxicity assessment end point.

The abundance and relevance of Accumulibacter phosphatis (presumed to be polyphosphate-accumulating organisms [PAOs]), Competibacter phosphatis (presumed to be glycogen-accumulating organisms [GAOs]), and tetrad-forming organisms (TFOs) to phosphorus removal performance at six full-scale enhanced biological phosphorus removal (EBPR) wastewater treatment plants were investigated. Coexistence of various levels of candidate PAOs and GAOs were found at these facilities. Accumulibacter were found to be 5 to 20% of the total bacterial population, and Competibacter were 0 to 20% of the total bacteria population. The TFO abundance varied from nondetectable to dominant. Anaerobic phosphorus (P) release to acetate uptake ratios (P rel /HAc up ) obtained from bench tests were correlated positively with the abundance ratio of Accumulibacter/(Competibacter 1TFOs) and negatively with the abundance of (Competibacter 1TFOs) for all plants except one, suggesting the relevance of these candidate organisms to EBPR processes. However, effluent phosphorus concentration, amount of phosphorus removed, and process stability in an EBPR system were not directly related to high PAO abundance or mutually exclusive with a high GAO fraction. The plant that had the lowest average effluent phosphorus and highest stability rating had the lowest P rel /HAc up and the most TFOs. Evaluation of full-scale EBPR performance data indicated that low effluent phosphorus concentration and high process stability are positively correlated with the influent readily biodegradable chemical oxygen demand-to-phosphorus ratio. A system-level carbon-distribution-based conceptual model is proposed for capturing the dynamic competition between PAOs and GAOs and their effect on an EBPR process, and the results from this study seem to support the model hypothesis. Water Environ. Res., 80, 688 (2008).