Asmaa Ali is a Research Data Scientist at Edelweiss Connect GmbH and a SaferWorldbyDesign Fellow, specialising in integrating artificial intelligence, data management, and toxicology to advance chemical risk assessment. With a solid academic foundation in Computer Science and Bioinformatics, she leverages her expertise in software development, machine learning, knowledge graphs, and data management to develop predictive models and innovative solutions for this critical field.
Currently, Asmaa is contributing to impactful projects, including toxicity prediction in aquatic species, data management and image analysis for nanomaterials characterization, and the application of language models to enhance chemical risk assessment practices.
Her professional journey spans several institutions. At the Egypt Center for Research and Regenerative Medicine, she optimised genomics pipelines and contributed to the Egyptian Genome Project. At Rosettastein Consulting GmbH, she developed machine learning models for chemical toxicity prediction. Asmaa has also made significant contributions to academia, serving as a teaching assistant for bioinformatics courses and leading AI-focused initiatives such as the OpenTox AI Hackathon 2023.
Passionate about the intersection of technology and science, Asmaa envisions a future where AI and precision toxicology converge to foster healthier lives and environments.
ORCID: https://orcid.org/0000-0001-9795-3489
Github: https://github.com/asmaa-a-abdelwahab
OpenTox 2.0 – A Perspective on the Principles for Predictive Toxicology and Risk Assessment Applications enabled by New Approach Methods, Computational Modelling and Artificial Intelligence
Asmaa Ali*, Daniel Burgwinkel, Barry Hardy (Edelweiss Connect)
*Presenter
In this presentation we review and discuss principles supporting the successful development, deployment and use of applications in predictive toxicology and risk assessment. We take a particular perspective based on the OpenTox framework and its principles originally formulated and published in 2010 (Hardy, B., Douglas, N., Helma, C. et al. Collaborative development of predictive toxicology applications. J Cheminform 2, 7 (2010). https://doi.org/10.1186/1758-2946-2-7. www.jcheminf.com/content/2/1/7).
We propose an updated set of principles, including ones for new applications based on Artificial Intelligence (AI), increased use of a variety of modelling and in vitro approaches often currently described as New Approach methods (NAMs), and their integration into integrated approaches such as Integrated Approaches to Testing and Assessment (IATAs) or Next Generation Risk Assessment (NGRA) or Safe and Sustainable by Design (SSbD) workflows. We review developments over the past fifteen years with regards to technical and scientific developments related to predictive toxicology and risk assessment. Many of the challenges and issues presented in the original OpenTox article (e.g., access to quality data, improved well-documented models, transparency, interoperability, ontology, applicability domain, validation, interpretation) remain valid today. We discuss the developments that have taken place over this period and also important challenges that remain. The main purpose of the presentation is to develop and communicate an updated well-formulated set of principles that can help guide effective application development and impact, particularly with a focus on open science approaches.
Note: This work is also being submitted as a full article for peer review in the special OpenTox issue of In Vitro Toxicology
