Not so fun fact: more than one million people die each year due to untreatable infections. Many of these COULD potentially have been saved with antibiotics, were it not for the fact that so called super bugs have evolved to be able to beat said treatments.
But it turns out that AI is now entering the arms race against super bugs....
Researchers are now turning to AI to discover new antibacterial drugs. At this stage of the game it's quite simple: feed the AI information about the structure of the drugs and their effectiveness at combatting nasties and the AI generates alternative structures which COULD be more effective.
The results have so far been pretty impressive. In one research project AI trimmed from a potentail 36 million compounds down to the ones that would kill harmful bacteria such as gonorrhoea and MRSA and it did this (obviously?) without harming human beings.
170 of the top candidates were then synthesized in the lab, and one of these cured mice of MRSA infections, and another cured gonorrhoea.
The results, reported in Cell, are a breakthrough. As lead author Professor James Collins told us, AI gives us "a leg up in the battle of our wits against the genes of superbugs."
New antibiotics are notoriously difficult and expensive to find. It can take more than ten years and billions of pounds for a new drug to make it to the market, and the majority of aspirants fail in the process. AI would change this by substantially accelerating the discovery process—screening through millions of possible molecules in silico before humans ever step into a trial.
IF these new products are deemed safe for human use, then we could see a new 'golden age' of antibiotics, but with a more rapid roll out as new resistances emerge, and who knows where that is going to end up....?
This just seems like a great way of spending a lot of money on making things that kill us better at killing us in the long term?
Or maybe I'm taking too simplistic a view...?
