Jean‐Michel Marin

The paper introduces an iterative framework for population genetic inference using simulation and machine learning, showing improved confidence interval coverage compared to existing methods.

This paper introduces TFscope, a machine learning method to identify genomic features explaining binding differences between TFs in ChIP-seq experiments, showing that co-factors and nucleotide environment often drive differences in cell types or treatments, while core motif differences are key for paralogous TFs.

The paper develops a sequence-based method to predict gene expression accurately, showing that nucleotide composition and regulatory region contributions are key factors.