Shengcheng Dong

The paper develops a scalable workflow to reclassify variants of uncertain significance using experimental and predictive evidence, resolving 75% of VUS in 40 genes.

The paper develops TLand, a machine-learning model that prioritizes and annotates non-coding regulatory variants by organ, using RegulomeDB features and outperforming existing models.

Explain-seq is a pipeline that trains and interprets sequence-based deep learning models to predict enhancer activities and identify transcription factor motifs in genomic data.

The paper updates RegulomeDB to improve non-coding variant annotation by integrating new functional genomic data and probabilistic scores from the SURF algorithm.

This paper introduces TURF, a tool to prioritize regulatory variants with tissue-specific function in non-coding regions using ENCODE data, showing high performance in predicting functional variants linked to GWAS traits.

The paper develops SEMpl, a pipeline that predicts SNP effects on transcription factor binding affinity using ChIP-seq data to identify disease-causing regulatory loci.