William J Greenleaf

Prospects for recurrent neural network models to learn RNA biophysics from high-throughput data

This paper evaluates a recurrent neural network model, Ribonet, for learning RNA biophysics from high-throughput data, showing it can predict RNA interactions and design functional RNAs but requires more diverse datasets for accurate biophysical modeling.

Structured nucleosome fingerprints enable high-resolution mapping of chromatin architecture within regulatory regions

The paper introduces NucleoATAC, an algorithm using ATAC-seq data to precisely map nucleosome positions and occupancy in yeast and human cells.