Ivan Ovcharenko

Using Deep Learning Models of Gene Regulation to Guide Drug Prioritization

This paper develops a deep learning framework linking noncoding genetic variation to drug repurposing by integrating enhancer prediction, gene expression changes, and drug-induced profiles.

Silencer variants are key drivers of gene upregulation in Alzheimer's disease.

This paper identifies silencer variants as key drivers of gene upregulation in Alzheimer's disease using a deep learning framework combining bulk and single-cell epigenomic data.

Comparative Analysis of Deep Learning Models for Predicting Causative Regulatory Variants

This study compares deep learning models for predicting causative regulatory variants using genomic and epigenomic data from multiple datasets, finding CNN models perform best for SNP regulatory impact prediction.

Super-silencers are crucial for development and carcinogenesis in B cells.

The paper identifies super-silencers in B cells, showing their role in development and their conversion to super-enhancers in lymphoma, with implications for carcinogenesis.

Regulatory plasticity of the human genome

The paper identifies enhancers prone to regulatory turnover using deep learning and shows they are repurposed between cell types, especially near neurodevelopmental genes.

Functional characteristics and computational model of abundant hyperactive loci in the human genome

The paper identifies and characterizes HOT loci in the human genome, showing they are widespread and formed by large transcriptional condensates, challenging the classical model of enhancer activity.

The contribution of silencer variants to human diseases.

The paper uses deep learning to analyze GWAS variants in silencer sequences across multiple tissues, finding strong disease associations and identifying specific SNPs linked to Parkinson's, schizophrenia, and Type 1 diabetes.

Modeling islet enhancers using deep learning identifies candidate causal variants at loci associated with T2D and glycemic traits.

Detection of new pioneer transcription factors as cell-type specific nucleosome binders

The paper develops a computational method to predict cell-type-specific nucleosome-binding abilities of transcription factors using ChIP-seq, MNase-seq, and DNase-seq data, identifying new potential pioneer transcription factors.

ChromDL: A Next-Generation Regulatory DNA Classifier

ChromDL is a deep learning model that improves the classification of regulatory DNA elements by combining BiGRU, CNN, and BiLSTM architectures, outperforming previous methods in TFBS, HM, and DHS detection.

De novo human brain enhancers created by single nucleotide mutations

The paper identifies ~4000 de novo human brain enhancers gained through single-nucleotide mutations, linked to neural development and CNS disorders.

A regulatory network of Sox and Six transcription factors initiate a cell fate transformation during hearing regeneration in adult zebrafish