​Our latest article, "Transfer Learning Enables Robust Prediction of Cellular Toxicity from Environmental Micro- and Nanoplastics," is now available online in the Journal of Hazardous Materials!

Micro- and nanoplastics (MNPs) are emerging pollutants of global concern; however, understanding their toxicity remains a significant challenge due to the limited data available. This work, led by postdoctoral researcher Dr. Kunpeng Chen, addresses a pressing challenge in environmental toxicology: the lack of robust toxicity prediction models for micro- and nanoplastics (MNPs).

In this study, we developed a transfer learning-based QSAR model that:
✔ Leverages pretraining on large nanoparticle datasets
✔ Fine-tunes on small MNP datasets with experimental and quantum chemistry features
✔ Outperforms conventional ML models in predicting cytotoxicity

✨ Key findings:
1. Surface functionalization and dose are critical toxicity predictors
2. Nanoplastics may induce cytotoxicity more rapidly than microplastics
3. The model shows strong potential for regulatory and risk assessment use

🔗 Read the full article here: https://lnkd.in/g2kemsJv

Kudos to Dr. Kunpeng Chen and all collaborators (Keyuan Li; Alexa Canchola; Qing Liu; Jinyong) on this impactful work!