AI, your age, and your bugs.

AI, your age, and your bugs.

THE GIST

Getting old isn’t only happening to you, it is happening to your personal microorganism community as well.   What we didn’t quite realize until as of late (thanks to a study out of UCSanDiego) is that you can actually predict the age of a person based on what’s going on with the critters in their bodies. How?  Using machine learning artificial intelligence (AI) technologies.    

THE BACKGROUND

The human body is host to millions of microorganisms, many of which are bacteria, and make up what is called the ‘microbiota’.  Most of this community of tiny life is acquired in the process of being born, and the first 5 years of life. Once they are with you – they are with you.  They help you digest, grow, and find many normal biological processes. What’s more, they exist in a delicate balance of good and evil. If the balance is disrupted, by diet or environment or a list of other excuses, they can cause disease.  It works the other way as well, disease can cause an imbalance potentially leading to other issues. For instance, Asthma is thought to be triggered by the introduction of certain bacteria into the microbiota. A recent study has found that Asthma also creates imbalances of microbiota in the airways alone, creating a greater bacterial burden than what exists in non-asthmatics.  Moral of the story, a healthy microbiota is worth keeping healthy and necessary for life. In that same theme, there is a lot to be revealed about a person just by peering into their micro-organism community.  

Why is this particular study remarkable?

Given the window that microbiota provides into the health of an individual, we are now able to view a different angle of the ageing process.  In the study, researchers sampled skin, mouth and fecal microorganisms. Guess which one yielded the most accurate results in predicting age? The skin! (Bet you thought it was poop).  Skin samples were able to deliver a prediction of the individual’s age – via machine learning – to within 3.8 years of accuracy. Machine learning? They pop their findings into a specifically programmed computer that uses AI to deliver the most accurate ageing prediction possible.  

Why was poop not more successful?  An interesting study from 2014 revealed that the gut microbiome is directly related to developmental maturity in infants – so if an infant is malnourished in any way, it will have a differently evolved gut biome than an infant who is not. These first few years are critical for that sort of development, and therefore easily influenced by external factors.  So some kids get a head start on their gut biomes in terms of ageing, while some might lag behind.  

The skin microbiota has an easier time.  It is influenced from a young age by environment and contact with its mother’s (or other caregiver’s) microbiota.  The changes that then occur to the skin are more predictable for the microbes to adapt to (ie: dry skin and shifts in oil production).  These guys can remain a little more consistent from the start of life to the end.    

When looking at the skin microbiota in the context of ageing, this level of accuracy to help us develop a better picture as to the overall health of an individual as they progress through life.  What’s more, it shows us yet another innovative application of machine learning and artificial intelligence – this time with the ability to look at the microbiome.

What does this mean for the future?

It means a couple of things. 

Firstly, as mentioned earlier,  we have a new perspective on ageing.  Ageing in many ways is still quite unknown to us.  As we start to paint a more complete picture, the element of microbiota is a vibrant addition.  This might help researchers look for diseases related to ageing from other angles. 

Second – the application of artificial intelligence and machine learning has now been proven to have significance when looking at microbiota.  This opens up a plethora of opportunities to create microbiome-based models that can scale and be applied in other contexts.