A Data Fusion Pipeline for Generating and Enriching Adverse Outcome Pathway Descriptions

Nymark, Penny; Rieswijk, Linda; Ehrhart, Friederike; Jeliazkova, Nina; Tsiliki, Georgia; Sarimveis, Haralambos; Evelo, Chris T.; Hongisto, Vesa; Kohonen, Pekka; Willighagen, Egon; Grafström, Roland C.

doi:10.1093/toxsci/kfx252

Cited by 53 publications

(46 citation statements)

References 55 publications

(90 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[ 3 ] Machine learning has been used extensively in nanoinformatics to develop predictive models for toxicity‐ and ecotoxicity‐related endpoints, employing various approaches such as read‐across methods, [ 4–6 ] nano‐quantitative structure–activity relationships (nanoQSAR [ 7–10 ] ), QSAR‐perturbation models, [ 11–13 ] and workflows predicting molecular initiating events and key events in adverse outcome pathways (AOPs). [ 14 ] Among the different types of descriptors used in predictive modeling approaches, image descriptors resulting from the analysis of electronic images of ENMs have been employed successfully. [ 15,16 ]…”

Section: Introductionmentioning

confidence: 99%

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

Karatzas

Melagraki

Ellis

et al. 2020

Small

Self Cite

View full text Add to dashboard Cite

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.

show abstract

Section: Introductionmentioning

confidence: 99%

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

Karatzas

Melagraki

Ellis

et al. 2020

Small

Self Cite

View full text Add to dashboard Cite

show abstract

“…Additionally, based on prior knowledge of interactions between genes and a predefined list of disease-associated markers, the pipeline for detecting AOP-linked molecular pathway descriptions has been described [57].…”

Section: Discussionmentioning

confidence: 99%

Transcriptional regulatory mechanisms of fibrosis development in mouse lung tissue exposed to carbon nanotubes

Zhernovkov

Santra

Cassidy

et al. 2018

Preprint

View full text Add to dashboard Cite

BackgroundCarbon nanotubes (CNTs) usage has rapidly increased in the last few decades due to their unique properties, exploited in various industrial and commercial products. Certain types of CNTs cause adverse health effects, including chronic inflammation and fibrosis. Despite the large number of in vitro and in vivo studies evaluating these effects, many important questions remain unanswered due to a lack of mechanistic understanding of how CNTs induce cellular stress responses. In order to predict CNT toxicity, it is important to understand which transcriptional programs are specifically activated in response to CNTs, and what similarities and differences exist in relation to other toxic inducers exerting similar adverse effects. ResultsA systems biology approach was applied to reveal complex interactions at the molecular level in mouse lung tissue in response to different fibrosis inducers: two types of multiwalled CNTs, NM-401 and NRCWE-26, and bleomycin (BLM). Based on mRNA gene expression profiles, we inferred gene regulatory networks (GRNs) to capture functional hierarchical regulatory structures between genes and their regulators. We found that activities of the transcription factors (TFs) Myc, Arid5a and Mxd1 were associated with the regulation of cytokine transcription in response to CNTs, while in response to BLM treatment, Myc was associated with p53 signaling. TF Litaf was identified as the essential regulator for noncanonical signaling of TLR2/4 driven by CNTs. Despite the different nature of the lung injury caused by CNTs and BLM, we identified common stress response modules, that included DNA damage (TFs : E2f8, E2f1, Foxm1), M1/M2 macrophage polarization (TF: Mafb), Interferon response (TFs: Irf7, Stat2 and Irf9) for all agents. ConclusionsThese results suggest that the reconstruction and analysis of TF-centric gene interaction networks can reveal key targets and regulators of cellular stress responses to toxic agents.2

show abstract

“…Therefore, an AOP can be used for all compounds where mechanistic evidence is available. AOPs can either be generated manually by literature review or mechanistic studies or by automated data mining . As mentioned earlier, high throughput screening data is a valuable tool to advance in silico toxicology.…”

Section: Adverse Outcome Pathwaysmentioning

confidence: 99%

In silico toxicology: From structure–activity relationships towards deep learning and adverse outcome pathways

Hemmerich

Ecker

2020

WIREs Comput Mol Sci

110

View full text Add to dashboard Cite

In silico toxicology is an emerging field. It gains increasing importance as research is aiming to decrease the use of animal experiments as suggested in the 3R principles by Russell and Burch. In silico toxicology is a means to identify hazards of compounds before synthesis, and thus in very early stages of drug development. For chemical industries, as well as regulatory agencies it can aid in gap‐filling and guide risk minimization strategies. Techniques such as structural alerts, read‐across, quantitative structure–activity relationship, machine learning, and deep learning allow to use in silico toxicology in many cases, some even when data is scarce. Especially the concept of adverse outcome pathways puts all techniques into a broader context and can elucidate predictions by mechanistic insights. This article is categorized under: Structure and Mechanism > Computational Biochemistry and Biophysics Data Science > Chemoinformatics

show abstract

A Data Fusion Pipeline for Generating and Enriching Adverse Outcome Pathway Descriptions

Cited by 53 publications

References 55 publications

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

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

Transcriptional regulatory mechanisms of fibrosis development in mouse lung tissue exposed to carbon nanotubes

In silico toxicology: From structure–activity relationships towards deep learning and adverse outcome pathways

Contact Info

Product

Resources

About