Related Experiment Video
Updated: Apr 16, 2026
09:11
Automation of the Micronucleus Assay Using Imaging Flow Cytometry and Artificial Intelligence
Published on: January 27, 2023
3.1K
In silico risk assessment for skin sensitization using artificial neural network analysis
Kyoko Tsujita-Inoue1, Tomomi Atobe, Morihiko Hirota
1Shiseido Research Center, Shiseido Co. Ltd.
The Journal of Toxicological Sciences
|March 20, 2015
Summary
Predicting chemical skin sensitization potential is crucial. An artificial neural network (ANN) model using in silico data shows promise, and combining it with in vitro data offers a powerful approach for integrated risk assessment.
Area of Science:
- Toxicology
- Computational Chemistry
Background:
- In vivo methods like the murine local lymph node assay (LLNA) are standard for assessing chemical skin sensitization.
- Ethical concerns and regulatory shifts necessitate the development of reliable alternatives to animal testing.
Purpose of the Study:
- To evaluate the predictive performance of artificial neural network (ANN) models using in silico descriptors for skin sensitization.
- To assess the combined utility of in vitro and in silico models for improved prediction of LLNA thresholds.
Main Methods:
- Development and application of an ANN model utilizing in silico-calculated 3D structural descriptors of chemicals.
- Integration of predictions from the in silico ANN model with a previously developed in vitro model (iSENS ver. 2).
Main Results:
- A good correlation was achieved between LLNA thresholds predicted by the in silico ANN model and reported experimental values.
- Combining in vitro and in silico model predictions reduced the under-estimation of chemical potency categories.
Conclusions:
- The ANN model employing in silico parameters demonstrates useful predictive performance for chemical skin sensitization.
- Integrating in silico and in vitro predictive models offers a promising strategy for the comprehensive risk assessment of skin sensitization potential.
