Artificial Intelligence Solves 120-Year-Old Biological Mystery


Computer scientists from Tufts University have published findings in PLOS Computational Biology detailing an unprecedented study in which the researchers wrote an artificial intelligence-based program to deconstruct a 120-year-old biological mystery until it had solved the puzzle. The most striking thing about the study, lead researcher Mark Levin says, was that the solution to the mystery “was not a hopelessly-tangled network that no human could actually understand, but a reasonably simple model that people can readily comprehend," suggesting that AI holds real promise in helping expedite scientific discoveries.

The mystery in question is that of the flatworm and its seemingly impossible regenerative abilities. A flatworm can be cut into many small pieces and each piece will regenerate, eventually growing back into many full-size flatworms each with complete sets of organs and anatomy. The regenerative properties of the flatworm’s DNA are a well-known phenomenon in science, called planaria. What is not known is how the process works at a genetic level. Regeneration has been studied fairly extensively in the past, but to date, researchers do not fully understand how genes coordinate the process to result in a specific final shape or anatomical entity. Levin explains, “You cannot tell if the outcome of many genetic pathway models will look like a tree, an octopus or a human. What we need are algorithmic or constructive models, which you could follow precisely and there would be no mystery or uncertainty. You follow the recipe and out comes the shape."

To tackle this problem, researchers created a program and seeded it with as much previous research findings as possible about the genetic properties of the flatworm, including previously discovered genetic functions involved in the regeneration process. Next, the program was given the genetic makeup of a dissected flatworm segment and tasked with constructing a genetic network that would allow the piece to regenerate, and that matched the findings of all previous studies on the subject. While this kind of data crunching would have taken lifetimes for humans to process, the program returned a viable result in three days. The regenerative properties of the flatworm, it found, are driven by two previously undiscovered proteins.

The results are important both for biologists and for researchers at large. Using the model, biologists expect to learn far more about the genetic instructions that drive regenerative properties in animals, while at a larger level, the study marks the first time in history that a computer, completely unaided by humans, was able to tackle a known scientific problem and return a valid working hypothesis.

Enjoy HIStalk Connect? Sign up for update alerts, or follow us at @HIStalkConnect.

↑ Back to top

Founding Sponsors

Platinum Sponsors