Programmers have developed an artificially intelligent algorithm, which they say can predict how old someone is with remarkable precision. All it needs is a sample of an individual's gut microbiome and it can narrow down their age to within four years, says a study available on the pre-print server bioRxiv, which is currently awaiting peer-review.
Your gut microbiome is a unique ecosystem consisting of trillions of microorganisms – some good, some bad, and some somewhere in-between. And we are only just starting to grasp how important it is to our health, our behavior, even how we process emotions like fear. In the past year alone, studies have been published linking gut flora to autoimmune diseases, degenerative disorders like Parkinson's, and autism. Another study examined the ways various gut microbiota may manipulate our moods via their waste products.
The latest study suggests there are consistent – and, therefore, predictable – ways the gut microbiome changes over a lifetime that make it possible for machine-learning software to accurately guess our age to within four years needing only a sample of our gut microbiome.
The algorithm used was designed by an artificial intelligence company called InSilico Medicine. Programmers used over 3,663 gut bacteria samples from 1,165 healthy individuals aged between 20 and 90 and from Austria, China, Denmark, France, Germany, Kazakhstan, Spain, Sweden, and the USA to train and then test their algorithm. Ninety percent of the samples were used in the training stage. The remaining 10 percent were used to test the algorithm's accuracy.
The result? Discounting a few anomalies, it could pinpoint the host's age to within 3.94 years, beating a previous model, which could predict a person's age with only 10 to 15 percent more accuracy that chance.
But that's not all. The machine was also able to work out which bacteria species were most important when determining a person's age. Apparently, out of hundreds of possible contenders, there are 39 species of gut flora that convey the most significant information when it comes to telling how old a person is.