My primary research interest is understanding genome regulation through the application of machine learning techniques. Both the complicated relationships between elements and the abundance of genomics data make machine learning, and in particular deep learning, well suited to this work. I actively develop and maintain software using these methods with the aim to improve traditional bioinformatics analysis, create new techniques, and make in silico predictions when bench work is infeasible or impossible. This includes LanceOtron, a tool developed by myself and colleagues, which applies a neural network to classify signal quality of chromatin profiling assays such as ATAC-seq, ChIP-seq, and DNase-seq.
GTAC enables parallel genotyping of multiple genomic loci with chromatin accessibility profiling in single cells.
Turkalj S. et al, (2023), Cell Stem Cell, 30, 722 - 740.e11
Direct correction of haemoglobin E β-thalassaemia using base editors.
Badat M. et al, (2023), Nat Commun, 14
LanceOtron: a deep learning peak caller for genome sequencing experiments.
Hentges LD. et al, (2022), Bioinformatics
Genotyping of Multiple Genomic Loci with Chromatin Accessibility Profiling in Single Cells Links Clonal Hierarchy with Epigenetic Variation in Acute Myeloid Leukemia
Turkalj S. et al, (2022), BLOOD, 140, 1193 - 1194
Defining genome architecture at base-pair resolution.
Hua P. et al, (2021), Nature, 595, 125 - 129