The Korean Society of Environmental Health and Toxicology (KoSEHT) is one of major scientific societies for Environmental Health in South Korea. They have 2 annual conference meetings (in Spring and Autumn). This year, the KoSEHT Spring Conference 2022 was held on 12-13 May 2022 in Busan, South Korea. This edition attracted more than 200 participants from academia, industry and regulators.

SABYDOMA’s partner KRICT (Dr Jongwoon Kim and Ms. Minju Na) represented the project in the national Korean conference. Under the title “Development of machine learning based nano-mixture QSAR models to predict mixture toxicity of metal oxides based nano-mixtures”, Minju presented a poster on her research based on previous studies on nano-mixture QSARs are very scarce and mainly focus on NP sizes smaller than 10 ~ 30 nm. Since those QSARs were developed based on density functional theory (DFT), having a limitation in calculating the descriptors of NP sizes larger than 10 nm mostly. Two machine learning algorithms (support vector machine (SVM) and Bayesian regularization for neural networks (BRNN)) were applied to develop models. Both calculated descriptor and all descriptor (including experimental properties) based models shows good performance with R2 > 0.7.

The fact that Metal oxide (MO) nanoparticles (NPs), which are widely used in various applications, have size-dependent unique physicochemical properties, means that toxicity can occur among chemicals at no-observed effect levels by the cocktail effect. As the nano-mixture toxicity is one of hot issues in the fields of toxicology and environmental science from the aspects of both sides on risk assessment and safe-by-design of mixture products, this work was very relevant for the conference.

Picture: Minju explaining her work to one of the participants of the conference