Hussein Mohsen

The paper introduces a dynamic clustering approach for genomics cohorts based on trait-specific genomic signals, revealing patterns beyond predefined racial or continental categories.

The paper identifies cancer-relevant genes by analyzing their proximity to core cancer genes in protein-protein interaction networks, finding that genes closer to cancer genes show greater functional importance in cancer biology.

The paper identifies genes with low mutation frequency but significant functional impact in 17 cancer types using network propagation, finding 8 previously unreported genes.

The paper investigates compression-based methods for interpreting deep learning models in cancer and psychiatric genomics, showing that compression quality can be assessed using generalization error and PAC-Bayes bounds.

SAUCIE is a deep neural network that processes large single-cell datasets, performs batch correction, and identifies cluster-based signatures in dengue patient T cells.

GRAM predicts the molecular effect of non-coding variants in a cell-type specific manner using SELEX and expression profiles, achieving high performance on MPRA and luciferase assays.