Scott S. Auerbach

Tox21mer is a transformer model trained on Tox21 concentration-response curves to predict compound activity and AC50, achieving high accuracy in outcome prediction and log10(AC50) estimation.

The paper predicts missing toxicogenomics data in the DrugMatrix database using ToxiCompl, a hybrid method combining matrix factorization and Bayesian optimization, showing alignment with biological expectations.

ALOHA improves dose-response analysis by using Bayesian splines to model gene responses and cluster co-regulated genes, leading to better identification of biological pathways and benchmark doses.

The paper develops an in-vitro gene expression screening method to predict chemical carcinogenicity and genotoxicity using HepG2 cells and machine learning models.