The growing problem of antibiotic resistance is a great threat to our global civilization and way of life.
It's almost unimaginable that we might return to a world without modern antibiotics, resorting to honey, sulfonamides, and tinctures of silver as we sufficed with in the 1930s.
The hunt is on for new compounds, biological or chemical, to save us from unmanageable bacterial infections.
Soil sampling processes have helped to find useful new microbes. This basically involves putting samples of mud in germ-ridden petri dishes to see if anything happens. By repeating this on a grand scale some promising new antimicrobial agents have been isolated. These new isolates can become potential new drugs of last resort to help keep the wolf from the door a little while longer.
However, the holy grail would be a whole new class of antibacterial agents that we have never seen before. This could buy us generations of time, potentially.
The ability for machine learning to make sense of out many complex variables has been put to great effect in this search. By feeding the system info on almost 2500 drugs and compounds known to have some effect on e. coli, they have been able to syncretise the X factor between them, and synthesise new drug concepts that would manifest this in a strong manner.
By focussing on an objective of 'kills germs but in a non-consensus way', this ML-driven system can hone in drugs that most scientists themselves would never consider trying to apply as antibiotics. Thus, out of chaos comes creativity once again, thanks to ML.
Extremely frustratingly, these promising new compounds may never see actual deployment any time soon. Pharma companies big and small are shying away from addressing these problems, due to an inability to make sufficient profit from antibiotics.
We have a tragedy of an overly-financialized commons, whereby no corporate leader wants to be the one to sacrifice short-term profit, even at the potential expense of an otherwise manageable future supergerm taking the life of one of their own children.
The incentive structures that once made contribution to society profitable are weakening, and we must reform them urgently, or suffer the consequences.