I came across this interesting article currently in peer review (thus may have issues):

Human microbiome aging clocks based on deep learning and tandem of permutation feature importance and accumulated local effects

If the results are correct, then it may give great insight to age-related diseases caused by the microbiome changing. The claim that a healthy’s person’s age can be determined with 4 years from their microbiome (with 95% accuracy) is a surprise, but very likely true.

“Some microbes showed steadily increasing age prediction with increasing abundance (e.g. [Eubacterium] hallii); other microbes were on the opposite, inversely correlating with predicted age (e.g. Bacteroides vulgatus)… certain microbes that were previously identified as important by PFI showed little influence on predicted age (e.g. [Eubacterium] rectale) ”

Among the interesting findings are the following that decreases with age:

• Bacteroides thetaiotaomicron
• Bacteroides vulgatus
• Bifidobacterium bifidum
• Bifidobacterium longum
• Odoribacter splanchnicus
• Streptococcus salivarius

The following increase with age:

• Lactobacillus reuteri
• Lactococcus lactis
• Propionibacterium freudenreichii

Bottom Line

Their AI model is not available on line (I wish it was and could accept ubiome data sets). It presents a significant step forward for determining what should be expected for an individual based on age. There are other factors that should also be included.

Addendum