Jerome F. Cros

The paper identifies distinct progression paths in pancreatic papillary cystic neoplasms using integrated transcriptomics and histology, enabling phenotype-specific prediction of high-grade dysplasia.

This paper benchmarks deconvolution algorithms for multi-omic data, identifying their strengths and limitations across different conditions and data types.

The paper develops a 16-gene signature (MG-GEM) that predicts survival benefit from adjuvant gemcitabine in resectable pancreatic cancer based on glycation stress, outperforming existing signatures.

This study shows that loss of HNF1B contributes to IPMN initiation and identifies its role as a tumor suppressor in pancreatic ductal cells.

DECONbench benchmarks deconvolution methods for quantifying tumor cell-type heterogeneity by providing datasets, tools, and performance evaluation.