Adrian J. Green

The paper develops and evaluates machine learning methods for transcriptome extrapolation in zebrafish using targeted gene sets, finding that neural network models outperform traditional methods.

This paper demonstrates that deep autoencoders outperform traditional statistical methods in detecting abnormal behavior in zebrafish exposed to toxicants.

The paper uses a conditional GAN and toxicity data to predict chemical toxicity, achieving higher sensitivity than a regression model but lower specificity.

This study shows that cadmium exposure disrupts otolith formation in zebrafish, leading to balance disorders and hyperactivity, with calcium rescue for otolith defects but not hyperactivity.