You are here:
Computational Intelligence: Opening DART's 'Black Box' with Agent-based Models
Citation:
Knudsen, T. Computational Intelligence: Opening DART's 'Black Box' with Agent-based Models. 11th World Congress on Alternatives and Animal Use in the Life Sciences Virtual Conference, NA, Virtual, NETHERLANDS, August 23 - September 02, 2021. https://doi.org/10.23645/epacomptox.16578704
Impact/Purpose:
Invited presentation to the 11th World Congress on Alternatives and Animal Use in the Life Sciences Virtual Conference August 2021 for symposium titled 'Computational Synthesis and Integration for Systems Toxicology in the Animal-free Zone'.
Description:
New Approach Methodologies (NAMs) for toxicity assessment that are based largely on in vitro data and in silico models provide a path forward to animal-free testing of the potential for developmental and reproductive toxicity (predictive DART). A major science challenge is translating complex data and information into predictive models for human toxicity. For predictive DART, this means virtually extrapolating data from in vitro studies with human stem cells and related systems into adverse developmental outcome(s) of relevance to regulatory decision-making. Computational agent-based models (ABM) of morphogenetic systems have distinct advantages for this purpose. ABMs offer unparalleled flexibility for multiscale modeling of tissue dynamics. Nature-inspired agents (cells) and rules (behaviors) are set into motion with soft-computing. Fuzzy logic is utilized to simulate forces or properties governing cell fate and behavior where rules are inexact or knowledge incomplete. ABMs can change course in response to a specific situation or stimulus from genetic and/or environmental cues from real world data such as in vitro high-throughput screening (HTS) yielding a probabilistic rendering of where, when and how a particular condition might lead to an adverse developmental outcome (cybermorphs). Opening the black-box of DART with computational intelligence comes with key challenges for science and technology development. Does not reflect Agency policy.
URLs/Downloads:
DOI: Computational Intelligence: Opening DART's 'Black Box' with Agent-based Models
WC11_KNUDSEN_852_V2.PDF (PDF, NA pp, 1820.355 KB, about PDF)