University of Toronto has launched at the end of July a spin off, Deep Genomics, that is leveraging on what once was called Artificial Intelligence and now is known as Deep Learning to mine the genome and find clues to diseases and cures.
Deep Learning, I wrote a few posts on that, is relatively new but at the core is the dream of having a machine that can demonstrate artificial intelligence and that can learn and get more knowledgeable (wiser?!). It is basically leveraging on software, in the form of neural networks, but more and more relies on special hardware, like the Synapse chip developed by IBM. Some prefer to keep AI separated from Deep Learning and indeed Deep Genomics claims to use both AI and DL.
The analyses of the genome is making use of this technology (in various hues...) to speed up sequencing and more and more to mine the secrets hidden into the many variations occurring in genomes.
The progress in genome sequencing has resulted in large data bases containing thousands and thousands of genomes and over 300 million variations that clearly would be impossible to study "manually".
These variations might be related, many are for sure, to diseases. Autisms, as an example, in some occurrences seems to have a genome alteration root cause. Intercepting the variations that matter is extremely complex and this is what Deep Genomics is set up to do.
They are developing tools, they already have one - SPIDEX- that can analyse various aspects of these variations. They also plan to offer these tools, and the access to the genomes data base to third parties to speed up understanding.
It is amazing to see software tools based on an intelligent paradigm exploring our code, that ultimately is at the root of our intelligence...