AbstractDecoding the cell type-specific transcription factor (TF) binding landscape at single-nucleotide resolution is crucial for understanding the regulatory mechanisms underlying many fundamental biological processes and human diseases. However, limits on time and resources restrict the high-resolution experimental measurements of TF binding profiles of all possible TF-cell type combinations. Previous computational approaches either can not distinguish the cell-context-dependent TF binding profiles across diverse cell types, or only provide a relatively low-resolution prediction. Here we present a novel deep learning approach, Leopard, for predicting TF-binding sites at single-nucleotide resolution, achieving the median area under receiver operating characteristic curve (AUROC) of 0.994. Our method substantially outperformed state-of-the-art methods Anchor and FactorNet, improving the performance by 19% and 27% respectively despite evaluated at a lower resolution. Meanwhile, by leveraging a many-to-many neural network architecture, Leopard features hundred-fold to thousand-fold speedup compared to current many-to-one machine learning methods.
Repression of a matrix metalloprotease gene by E1A correlates with its ability to bind to cell type-specific transcription factor AP-2.
