Deep Learning on Neanderthal Genes

This is the seventh post of my column Deep Learning for Life Sciences where I give concrete examples of how Deep Learning can already now be applied in Computational Biology, Genetics and Bioinformatics. In the previous posts, I demonstrated how to use Deep Learning for Ancient DNA, Single Cell Biology, OMICs Data Integration, Clinical Diagnostics and Microscopy Imaging. Today we are going to dive into the exciting History of Human Evolution and learn that it is straightforward to borrow methodology from the Natural Language Processing (NLP) and apply it to Human Population Genetics in order to infer regions of Neanderthal introgression in modern human genomes. When ancestors of Modern Humans migrated out of Africa 50 000 - 70 000 years ago, they encountered Neanderthals and Denisovans, two groups of ancient hominins that populated Europe and Asia at that time. We know that Modern Humans interbred with both Neanderthals and Denisovans since there is evidence of the presence of their DNA in genomes of Modern Humans of non-African origin.