Harmful algal blooms (HABs) are a rising health and environmental concern in the United States, particularly in South Florida. Skin contact and the ingestion of contaminated water or fish and other seafood have been proven to have severe toxicity to humans in some cases. However, the impact of aerosolized HAB toxins is poorly understood. In particular, knowledge regarding either the immediate or long-term effects of exposure to aerosolized cyanotoxins produced by freshwater blue-green algae does not exist. The aim of this study was to probe the toxicity of aerosolized cyanobacterial blooms using Drosophila melanogaster as an animal model. The exposure of aerosolized HABs at an early age leads to the most severe long-term impact on health and longevity among all age groups. Young groups and old males showed a strong acute response to HAB exposure. In addition, brain morphological analysis using fluorescence imaging reveals significant indications of brain degeneration in females exposed to aerosolized HABs in early or late stages. These results indicate that one-time exposure to aerosolized HAB particles causes a significant health risk, both immediately and in the long-term. Interestingly, age at the time of exposure plays an important role in the specific nature of the impact of aerosol HABs. As BMAA and microcystin have been found to be the significant toxins in cyanobacteria, the concentration of both toxins in the water and aerosols was examined. BMAA and microcystin are consistently detected in HAB waters, although their concentrations do not always correlate with the severity of the health impact, suggesting the potential contribution from additional toxins present in the aerosolized HAB. This study demonstrates, for the first time, the health risk of exposure to aerosolized HAB, and further highlights the critical need and importance of understanding the toxicity of aerosolized cyanobacteria HAB particles and determining the immediate and long-term health impacts of HAB exposure.

Deep learning is a machine learning technique to model high-level abstractions in data by utilizing a graph composed of multiple processing layers that experience various linear and non-linear transformations. This technique has been shown to perform well for applications in drug discovery, utilizing structural features of small molecules to predict activity. Here, we report a large-scale study to predict the activity of small molecules across the human kinome—a major family of drug targets, particularly in anti-cancer agents. While small-molecule kinase inhibitors exhibit impressive clinical efficacy in several different diseases, resistance often arises through adaptive kinome reprogramming or subpopulation diversity. Polypharmacology and combination therapies offer potential therapeutic strategies for patients with resistant diseases. Their development would benefit from a more comprehensive and dense knowledge of small-molecule inhibition across the human kinome. Leveraging over 650,000 bioactivity annotations for more than 300,000 small molecules, we evaluated multiple machine learning methods to predict the small-molecule inhibition of 342 kinases across the human kinome. Our results demonstrated that multi-task deep neural networks outperformed classical single-task methods, offering the potential for conducting large-scale virtual screening, predicting activity profiles, and bridging the gaps in the available data.