Description
Recently, Hi-C has been used to probe the 3D chromatin architecture of multiple organisms and cell types. A computational method for predicting pairwise contacts without the need to run a Hi-C experiment would be invaluable in understanding the role that 3D chromatin architecture plays in genome biology. In this talk I will describe Rambutan, a Python package that uses a deep convolutional neural network to predict Hi-C contacts at 1 kb resolution using nucleotide sequence and DNaseI assay signal as inputs. I will explain the area of chromatin architecture briefly, how Rambutan fits in with other methods in the field, validate this model using both traditional accuracy metrics and at biological tasks, and show how one can use Rambutan simply in their own work.