Evan Paull

DoFormer improves perturbation prediction by using a causal Transformer that distinguishes observational from interventional regimes without DAG assumptions.

GREmLN is a foundation model that uses graph structure to improve gene embeddings in single-cell transcriptomics, outperforming baselines in cell type annotation and perturbation prediction.

The paper identifies treatment-resistant Cancer Stem-Like Cells in breast cancer using single-cell RNA sequencing and validates a drug combination to target them effectively.

The paper uses CRISPRi and single-cell RNA-seq to reverse engineer gene regulatory networks in colorectal cancer cells, revealing autoregulatory and cooperative mechanisms.

The paper identifies 407 Master-Regulator proteins that canalize cancer genomics into transcriptional subtypes and links genetic alterations to tumor biology through modular regulatory architectures.

The paper introduces an algorithm to reconstruct context-specific molecular interaction maps centered on oncoproteins, validating it with KRAS in lung adenocarcinoma and identifying novel synthetic lethal proteins.

The paper identifies four molecularly distinct glioblastoma subtypes using single-cell protein-activity analysis and shows drug sensitivity is state-dependent.