The world has been transformed by antibiotics.
They’ve helped combat diseases that would previously be anything from discomfort to death. But, no new types of antibiotics have hit the market since the 1980s.
What if the oldest, most distant relatives held the key to overcoming resistance to antibiotics?
Some scientists, like University of Pennsylvania bioengineering professor Cesar de la Fuente would like to discover new antibiotics by using machines learning … as well as a few very, very old cousins.
Machines and molecular innovations
The world has been transformed by antibiotics and made it possible to treat illnesses which used to be a sign of everything from discomfort to death.
Now, we face an entirely new challenge.
“We’re confronted with a silent pandemic in which more and more bacteria are becoming resistant to the available antibiotics” de la Fuente states.
As a postdoctoral researcher in MIT, de la Fuente was struck by the idea of what would it be like if machine learning could show the computer to invent on a molecular scale?
His team achieved exactly what they wanted — they trained computers to implement Darwin’s algorithm for evolution. The team in 2018 released the first study to their knowledge. their initial study using AI to identify a novel antibiotic.
“It used antibiotics at first which were not particularly effective, and was able to evolve into being more efficient,” he says. The modern antibiotics destroyed bacteria that were in mice.
Mining proteins inherited from our ancestral ancestors
Then, de la Fuente and his colleagues used the computer models to scour the proteins found in the human body, referred to as the proteome, in search of small proteins called peptides which could play a part for the immune system.
They have discovered more than 2500 peptides that have anti-infective properties They wondered: What could they do if they shifted their focus towards the extinct species as they hunt for novel antibiotic molecules?
De la Fuente says organismal de-extinction, a concept from Jurassic Park, kept coming to mind during discussions. However, instead of dinosaurs they focused on the closest relatives of humans Neanderthals as well as Denisovans.
“Instead in bringing back whole organisms, why don’t we just bring back the molecules of the past to help solve our the current problems?” de la Fuente declares.
De la Fuente says he together with his team that – they created the machine-learning model which could extract proteomic as well as genomic information from Neanderthals and Denisovans. The model identifies sequences from archaic human beings and suggests the ones that are likely to be effective antibiotic candidates for antibiotics.
Next step? Resurrection.
“We employ a method known as chemical synthesis in solid phase, which is basically a set of robots that enable us to create peptides. They make just one amino acid at a given time, and then connect them together into a chain to create a peptide which is in reality small in size,” de la Fuente describes. “And then, we expose them to the bacteria we cultivate in the lab and test whether they are able to eliminate the bacteria that are clinically relevant or not.”
They discovered a variety of peptides that effectively destroyed organisms in petri dishes and then tested them on animal models.
“In some of these mouse models that was a skin-infection model one of Neanderthal protein peptides were able to lower the severity of the infection to levels that were comparable to a regular antibiotic, known as Polymyxin B,” de la Fuente declares.
They referred to it “neanderthalin-1” as well. even though the peptide itself isn’t powerful enough to function as an antibiotic in its own, de la Fuente says the team hopes to utilize it along with other peptides as models to further research anti-microbials.
Want more on de-extinction? We’ve got your back! Watch our show about the demise of whole species, including the woolly mammoth and the dodo.
Have a question? Contact us at shortwave@npr.org.
Enjoy Short Wave on Spotify, Apple Podcasts and Google Podcasts.
The episode was created by Rachel Carlson. The editing was done by Rebecca Ramirez. The fact checker was anil Oza. The Audio engineers were Patrick Murray.
9(MDEwMjQ0ODM1MDEzNDk4MTEzNjU3NTRhYg004))
