Deep Learning Essentials

Genomics

One of the interesting application areas of deep learning is genomics, where advanced CNN models are used to learn structure from large and high-dimensional DNA datasets. One of the earliest applications in this space was using handcrafted features with the full connected feed forward neural network for predicting the splicing activity of an exon. Recently, a new technique has been proposed in the form of an open source implementation called Basset: Learning the regulatory code of the accessible genome with deep convolutional neural networks (http://genome.cshlp.org/content/early/2016/05/03/gr.200535.115.abstract), which predicts gene accessibility (DNase I hypersensitivity) across 164 cell types. To solve this problem, Basset uses ...