Computers looking for new antibiotics
2020/03/03 Roa Zubia, Guillermo - Elhuyar Zientzia
.jpg)
Artificial intelligence is a great sack. It is, in short, computer techniques for the realization of very complex works, which are generally characterized by computer learning. Computer scientists train the computer and when it is formed performs complex tasks. Look for antibiotics, for example.
Looking for antibiotics is not easy. It is a very complex work. It is somehow a war against bacteria. Antibiotics are molecules that kill bacteria. Penicillin or streptomycin do this. Some bacteria are killed. The problem is that over time bacteria learn how to cope with antibiotics. Resistant. And therefore, humans need to find new antibiotics to kill bacteria.
For the first time they have looked for new antibiotics using artificial intelligence, teaching a computer to look for them. And they find a new one, an antibiotic. He has paid homage to the computer HAL9000, which appears in the novel 2001: the Witch.
The computer has learned to look for antibiotics by analyzing 2,300 molecules that block the growth of the bacterium Esclerichia coli and, once this study is completed, experts have asked you to announce which other molecule can be effective against bacteria. It has given many results, many possible antibiotics. But they had to be tested. Experts have tested 100 of them, one of which has given very good results with clinical sessions in mouse.
Perhaps the most interesting thing is not that a computer did. The most interesting thing is that the computer works in a very different way than the human. Scientists study how the bacteria grow and test molecules that can break that growth. Instead, the computer looks for the potential effects of each molecule, and if this effect breaks the growth of the bacteria, it proposes to be a candidate for antibiotics. Reverse route. In fact, the operation of the beacon seems not to be very effective: it blocks protons traversing the bacterial membrane, which is supposedly not a very destructive strategy, but in clinical trials it has had a good result against many highly resistant bacteria.
It's a pretty story, but the alicine will also have its dark side. Although it is now effective, over time bacteria can learn to block it. The problem of resistance has not been solved, but for now it is good news that artificial intelligence has found a new antibiotic. We have a tool and will use it.
