Deep Learning in Bioinformatics: a Review

Dimitri Nowicki (Associate research professor at Institute of Cybernetics of NASU)

Abstract: We will present a review of fresh results in deep learning for bioinformatics.
First, we wil give a short introduction to bioinformatics, nature of its data and main problems.
Then we will briefly describe methods of mapping from Levenstein to Euclidean metrics, and unsupervised learning and clustering of bioinformatic data.
Next, we consider application of deep neural networks for protein function prediction as well as protein folding

Slides

Deep Spatio-Temporal Architectures and Learning for Protein Structure Prediction

Machine Learning: An Indispensable Tool in Bioinformatics

Deep Learning for mining big data in the natural sciences

Principal component analysis for clustering gene expression data

Reconstructing Protein Structures by Neural Network Pairwise Interaction Fields and Iterative Decoy Set Construction