Details

Autor(en) / Beteiligte

• Karatzas, Pantelis
• Melagraki, Georgia
• Ellis, Laura‐Jayne A.
• Lynch, Iseult
• Varsou, Dimitra‐Danai
• Afantitis, Antreas
• Tsoumanis, Andreas
• Doganis, Philip
• Sarimveis, Haralambos

Titel

Development of Deep Learning Models for Predicting the Effects of Exposure to Engineered Nanomaterials on Daphnia magna

Ist Teil von

• Small (Weinheim an der Bergstrasse, Germany), 2020-09, Vol.16 (36), p.e2001080-n/a

Ort / Verlag

Germany: Wiley Subscription Services, Inc

Erscheinungsjahr

2020

Quelle

Wiley Online Library Journals Frontfile Complete

Beschreibungen/Notizen

• This study presents the results of applying deep learning methodologies within the ecotoxicology field, with the objective of training predictive models that can support hazard assessment and eventually the design of safer engineered nanomaterials (ENMs). A workflow applying two different deep learning architectures on microscopic images of Daphnia magna is proposed that can automatically detect possible malformations, such as effects on the length of the tail, and the overall size, and uncommon lipid concentrations and lipid deposit shapes, which are due to direct or parental exposure to ENMs. Next, classification models assign specific objects (heart, abdomen/claw) to classes that depend on lipid densities and compare the results with controls. The models are statistically validated in terms of their prediction accuracy on external D. magna images and illustrate that deep learning technologies can be useful in the nanoinformatics field, because they can automate time‐consuming manual procedures, accelerate the investigation of adverse effects of ENMs, and facilitate the process of designing safer nanostructures. It may even be possible in the future to predict impacts on subsequent generations from images of parental exposure, reducing the time and cost involved in long‐term reproductive toxicity assays over multiple generations. This work presents a workflow for detecting automatic malformations in Daphnia magna images using predictive models, developed by applying deep learning methodologies to a big collection of images. An object detection model isolates the regions of interest and two classification models rank specific objects (abdomen/claw, heart) in terms of lipid concentrations. Comparison with control images can indicate the types and levels of malformations.