We have to applaud the way that the UK Government tries to take questions from its citizens as part of its response to the Covid-19 pandemic. No taxation without consultation, one might say.
The question was “Will the UK Government fund research into Artificial Intelligence algorithms, thereby helping it to predict the next mutation in the SARS-CoV-2 virus?
Alright, they take a lot of questions, and we didn’t make the cut! But for those who missed our last outing on this theme (LSS; 28April; Using AI …..) here is a brief summary of our ideas.
All genomes, that is to say the nucleic acids strands in our cells, have hotspots. These are places where the normal rate of mutation seems to speed up. According to Patricia Frost in Science Daily the mutations are 18 times more likely in DNA sequences where the same chemical letter, such as A, G, C, or T is repeated multiple times. They are 12 times more likely where you find repeated triplets, e.g, TTT or AAA.
Now, we know more astute readers will protest that coronaviruses code in RNA. Yet this is where the fun starts. RNA viruses are more prone to mutation, as they are dependent on certain error-prone enzymes, and have less in built correction mechanisms, than more advanced DNA -based organisms.
In theory, it should be possible to produce a model of the viral genome, and to map its mutation hotspots. More importantly, the possible mutational pathways of the virus should be predictable.
Work to analyse viral mutations is already been undertaken for influenza, though we don’t know what computational techniques the authors used (see link below)
A simple survey of Wikipedia demonstrates the enormous power and range of AI applications, though we could not resist adding a post from fellow blogger Aayushi Johari. Why not now deploy this power on an urgent and pressing problem, instead of designing tired old games, that frankly, everyone has seen before? A success here would be national prestige indeed.
#sarscov2 #covid19 #borisjohnson #genome #mutation #artificialintelligence