Remo Rohs

This paper introduces a model-guided virtual screening pipeline to improve synthesizability in generative drug design by combining SBDD models with chemical similarity search.

The paper investigates DNA breathing's role in TF-DNA interactions using in vitro and in vivo data, finding correlations between breathing features near motifs and binding affinity across multiple TFs.

DeepPBS predicts protein-DNA binding specificity using geometric deep learning, showing alignment with experimental mutagenesis data and enabling analysis of molecular dynamics simulations.

DrugHIVE is a deep generative model that accelerates drug design tasks by enabling fine-grained molecular generation and outperforming existing methods in speed and performance.

The paper investigates how DNA sequence and chromatin context influence transcription factor binding in Plasmodium falciparum, showing that both factors contribute to differential binding site selection.

PNAbind predicts protein-nucleic acid binding using graph neural networks, achieving high accuracy in binding site prediction and discrimination between DNA and RNA specificity.

This paper identifies a cis-regulatory mechanism in triple-negative breast cancer where targeting circadian transcription factors with small molecules represses a circadian transcriptional program, offering a new therapeutic strategy.

This paper introduces Deep DNAshape, a deep learning method that predicts DNA shape by considering extended flanking regions, improving prediction accuracy and providing insights into protein-DNA binding mechanisms.

This study identifies Hand2's role in repressing non-cardiac cell fates through chromatin remodeling at cis-regulatory elements in zebrafish cardiac progenitors.

The paper quantifies the spatial activity of a Drosophila enhancer through sequence mutagenesis and image analysis, revealing dense regulatory information and novel regulatory logic.

The paper introduces a sequence+shape kernel to model transcription factor binding without requiring aligned binding sites, showing improved prediction of binding affinity across different k values.