Rachel Melamed

This paper introduces a transformer-based model, DraPhormer, to explain drug therapeutic and adverse effects by integrating drug molecular data, disease genetics, and known drug effects.

This paper disentangles disease genes contributing to drug therapeutic benefits versus side effects using a network model and identifies insights for drug repurposing and treatment selection.

This paper uses genetic similarity between diseases to predict drug therapeutic and side effects, finding higher genetic similarity correlates with shared drug outcomes.

The paper imputes gene expression in 103 brain tissues using BRONTE, a graph neural network, and identifies new genes and tissues affected by genetic variants linked to brain-related traits.

The paper identifies novel breast cancer treatment candidates by leveraging shared genetic associations between breast cancer and its predisposing diseases.

The paper uses Mendelian disease comorbidity to identify potential drug targets for complex diseases by linking causal genes and their associated drugs.

This paper develops Draphnet, a model that connects drug molecular effects to disease genetics to predict and explain drug impacts on disease phenotypes.

This paper develops indication embeddings to reduce health data dimensionality, improving cohort matching for drug safety studies by capturing relevant treatment decision information.

The paper analyzes prescription patterns of 600 drugs across 2,334 US counties using medical claims data, revealing geographic disparities in treatment choices and drug preferences.