Matthew E. Gosden

This study identifies the transcriptional regulatory mechanisms governing ERFE expression in human erythroblasts using genomic approaches.

The paper identifies a de novo enhancer mutation in KCNB1 causing ST Depression Syndrome, a cardiac arrhythmia, and shows it can be predicted by a deep neural network.

The paper shows that super-enhancers function directionally and orientation-dependently, unlike regular enhancers, by analyzing the β-globin super-enhancer.

This paper establishes a scalable protocol to isolate primitive-like erythroid cells from in vitro differentiated mESCs, showing they recapitulate normal erythropoiesis and mutation effects seen in primary cells.

DeepC predicts chromatin interactions from DNA sequence using deep learning and transfer learning, generalizing across cell types and resolutions.

The paper shows that actively transcribed genes can function as chromatin boundary elements, similar to CTCF-bound sites, by stalling chromatin loop extrusion.

The paper studies chromatin architecture changes during mouse erythroid differentiation using Tiled-C and shows enhancer-promoter interactions form during differentiation, not before gene expression.

The paper introduces a new method, NuTi Capture-C, that improves high-resolution chromosome conformation capture by reducing noise and increasing sequencing content, enabling reproducible genome-wide interaction profiles.

The paper develops a platform combining ATAC-seq, DNaseI footprints, NG Capture-C, and machine learning to identify causal variants and genes for polygenic human traits.

This study models erythropoiesis in CDA-I by analyzing patient-derived erythroblasts, identifying abnormalities in DNA accessibility, proliferation, and enucleation, with genotype-phenotype correlations.