Hyun Jae Cho

This study develops a variational autoencoder to generate synthetic scRNA-seq data from bulk RNA-seq, enabling single-cell insights from large-scale bulk datasets.

DARDN uses deep learning to classify cancer-specific CTCF binding sites and identify oncogenic transcription factors in multiple cancer types.

Bedshift generates perturbed BED files to benchmark region set similarity metrics, showing differences in their sensitivity to added, dropped, or shifted regions.

The paper introduces a method to represent genomic region sets as embeddings using a modified word2vec approach, showing it effectively captures biological associations in lower dimensions.