Event Calendar
Prev MonthPrev Month Next MonthNext Month
ASCCT-ESTIV Award Winners Series: Dr. Eliska Kuchovska & Tiago Marques Pedro
Friday, November 22, 2024, 10:00 AM - 11:00 AM EST
Category: ASCCT Webinar
REGISTER NOW

Featuring:

Dr. Eliska Kuchovska, PhD: “Ontology-Based AI-driven Innovative Approach Using DNT NAMs for NGRA.”  JSAAE Best Oral Presentation award at ESTIV 2024

Tiago Marques Pedro: “Using a Machine Learning Framework to Improve the Efficiency of Mitochondrial Toxicity Screening by Guiding Compound Selection.”  ESTIV Best Oral Presentation award at EUROTOX 2024

A brief Q&A session will follow each presentation.

ABSTRACTS

Ontology-Based AI-driven Innovative Approach Using DNT NAMs for NGRA: The current regulatory guidelines for assessing developmental neurotoxicity (DNT) are inadequate for hazard assessment of the vast array of chemicals in our environment. Additionally, our understanding of human brain development often relies on animal-derived data, which may not accurately represent the human context. Therefore, there is a critical need for more reliable and efficient human-centered new approach methodologies (NAMs), ideally combining in vitro and in silico methods: a strategy applied in the European H2020 ONTOX project [1].

To construct the DNT ontology, a framework for data integration, we first shaped its pillars. The first pillar consists of a comprehensive physiological map of the developing human brain, serving as a foundational knowledge base that provides insights into the biological underpinnings and physiological mechanisms crucial for healthy brain development. The second pillar is a meticulously curated adverse outcome pathway (AOP) network which describes events leading to the adverse outcome of decreased cognitive function and impaired learning and memory. Subsequently, a tailored human in vitro battery was selected and characterized regarding its biological applicability domain to determine its relevance to human physiology and neurodevelopmental disorders.

The DNT in vitro NAMs, as well as the physiological map, also serve to derive new AOPs to complete the current AOP network. Furthermore, the DNT ontology and assay characterization contribute to bolstering the regulatory acceptance of in vitro NAMs by reducing their uncertainty. Ultimately, the final iteration of this ontology-based approach, which combines in silico and in vitro NAMs with exposure assessment, will predict the effects of chemicals on the developing human brain without the need for animal testing and advance human risk assessment in line with the principles of next generation risk assessment (NGRA).

[1] Vinken, M., Benfenati, E., Busquet, F., Castell, J., Clevert, D.-A., de Kok, T. M., Dirven, H., Fritsche, E., Geris, L., Gozalbes, R., Hartung, T., Jennen, D., Jover, R., Kandarova, H., Kramer, N., Krul, C., Luechtefeld, T., Masereeuw, R., Roggen, E., … Piersma, A. H. 2021, ‘Safer chemicals using less animals: kick-off of the European ONTOX project‘, Toxicology, 458, 152846

Using a Machine Learning Framework to Improve the Efficiency of Mitochondrial Toxicity Screening by Guiding Compound Selection: Mitochondrial dysfunction plays a major role in the onset of off-target drug effects, including hepatotoxicity and cardiotoxicity. Early identification of potential mitochondrial toxicants during drug development is essential to prevent these off-target toxicities. Despite the utilisation of established assays for mitochondrial function measurement, these methods usually lack efficiency and intent, reflected in their indiscriminate approach to compound screening selection. Therefore, harnessing machine learning (ML) workflows may provide avenues to improve the screening proficiency of mitochondrial toxicants.

Active learning (AL), a ML framework, was employed to guide compound selection and improve screening time to demonstrate the strengths of this hybrid screening approach. An initial screen, conducted on 1520 compounds from the Prestwick Chemical Library using an ATP cell viability assay in metabolically-switched HepG2 cells, revealed over 100 mitochondrial toxins. These results were then used to iteratively train the AL model, demonstrating a 2-fold improvement in identifying true positives compared to random selection when only half of the library was used in the training set. Further investigation is required to determine the features driving uncertainty, with structural similarity across mitotox class, physicochemical properties and mechanism of action theorised as potential key players. This study demonstrates integrating AL-enhanced mitochondrial toxicity screening may reduce screening time whilst preserving true positive identification for drug safety assessment.

About the Presenters

Eliška Kuchovská is a postdoctoral researcher at the Leibniz Research Institute for Environmental Medicine in Germany, specializing in the development of human-relevant, animal-free New Approach Methodologies (NAMs) to evaluate the effects of chemicals on brain development. Driven by her fascination with the complexity of the brain, Eliška combines rigorous research with an enthusiasm for science communication. She actively promotes career development for early-career researchers, serving as chair of the ASPIS Academy and the EUROTOX Early Career Forum.  

Tiago Marques Pedro earned his BSc in Biomedical Science at the University of Surrey, UK, which included a research placement year in toxicology at Weill Cornell Medicine in New York, USA. He went on to pursue a Master's in Toxicology at Karolinska Institute in Stockholm, Sweden, where he explored the effects of environmental pollution mixtures on genotoxicity, and the use of DNA repair inhibitors to combat resistance to cisplatin-based chemotherapies, across various research projects. Currently a 3rd year PhD student at the University of Cambridge, Tiago's project focuses on using machine learning to predict mitochondrial toxicity.

 

Registration: https://us06web.zoom.us/webinar/register/WN_ZdJ5Lj5DSiKmYO2y_B1FNQ

Recordings and other materials from this webinar will be posted on the ASCCT webinar archive: https://www.ASCCTox.org/Webinar-Archive