Drug-resistant bacteria are linked to millions of deaths a year and make once-effective treatments useless. Beating resistance would keep bacterial infections treatable and could avert many deaths in the coming decades.
What we know
Resistant bacteria are associated with millions of deaths a year worldwide.
What blocks us
New antibiotics earn little, which has limited the interest of investors and most large drug companies.
Current leads
AI-discovered antibiotics such as halicin, and bacteria-killing viruses (phages).
Sub-questions
How can AI models that search for new antibiotics be made reliable outside test conditions and understandable rather than 'black boxes'?
How can we predict how a new drug candidate will behave inside the human body when so little data exist?
Can we predict whether a bacteria-killing virus (phage) will destroy its host right away or hide inside it?