Title:

Comparative analysis of methods for assessing the irritant and allergic potentia! of cosmetic ingredients in the context of the ban on animal testing

Author:

Kalicińska, Jadwiga

Subject:

cosmetics ; irritant potential ; contact allergy ; risk assessment ; artificial intelligence

Abstract:

In 2004, a ban on testing cosmetic products on animals has come into force in the EU, however, the question arises whether in vitro methods reliably reflect processes occurring in the human body. The aim of this study was I to compare the allowed in vitro methods with historical I data and to develop an in silico model as a possible altemative. Based on the analysis of the literature, a list of irritating or sensitizing cosmetic ingredients was established and compared with the published results of I I human, animal and in vitro studies. Using WEKA software, an in silico model was created to assess the sensitization potentia! of cosmetic ingredients based on physicochemical properties. The model was trained using epidemiological data compiled into indices of eOR ( estimated odds ratio) and eAR ( estimated attributable risk) for 78 ingredients. Among the tested in silico models, the best results were obtained with the Naive Bayes model based on the eAR index and trained with 9 descriptors - its accuracy was 86%, sensitivity 80%, and specificity 90%. This model can also be used in the evaluation of drugs for extemal use. Conclusion: The developed in silico model allows to predict the sensitizing potential of cosmetic ingredients with an accuracy comparable to historical results of animal tests and currently used in vitro tests, yet at a much lower cost.

Place of publishing:

Kraków

Level of degree:

2 - studia doktoranckie

Degree discipline:

dermatologia

Degree grantor:

Rada Dyscypliny Nauki farmaceutyczne

Promoter:

Śpiewak, Radosław ; Wiśniowska, Barbara

Date:

2023

Date issued:

2023

Type:

Praca doktorska

Call number:

ZB-139420

Language:

pol; eng

Access rights:

nieograniczony

×

Citation

Citation style: