Enabling faster drug discovery through machine learning

The goal of my project is to develop and test algorithms to speed up the drug discovery process. I will go about this in a number of ways. The first method, to speed up the process, is to develop algorithms that can specifically identify drugs to fight a given disease. As such, the input to the algorithm will be certain information about the disease, such as known binding receptors, and the output will be a group of molecules that will aid in combating the disease. The second method will be to develop algorithms that can optimise the retro-synthetic process. If a chemist is given a compound that they wish to create, retro-synthesis is the process of calculating how the molecule can be formed from other compounds that are readily available. There are many possible routes to a given compound and the optimisation of this process, from a computational perspective, could really save a lot of time. All this work will be in collaboration with GSK, I will design and
test the algorithms whilst they produce the drugs in their labs.