Researchers have evaluated the ability of the oral, gut and skin (palm and forehead) microbiomes in predicting age in adults. From this, they have identified that the skin is the best predictor of age.
Evidence has shown that environmental factors influence the composition of our gut microbiota which changes as we age. The microbiome rapidly changes in the first three years of life and then relatively little in adulthood. Post-mortem studies have revealed that the microbiome continues to change after death. Interestingly, the skin microbiome has been able to predict post-mortem interval much better than the gut.
In a study, published in the American Society for Microbiology, researchers expanded on previous work on age prediction in the gut microbiome to other body sites. They used a total of 4,434 faecal, 2,550 saliva and 1,975 skin samples from various countries across the world.
Machine learning algorithm
The team used an algorithm (random forests) to regress relative abundances of amplicon sequence variants (ASVs) in the healthy human microbiota from the different body sites against the participants’ chronological ages.
The researchers regression analysis supported previous evidence that associated the gut microbiome with chronological ageing. Nevertheless, the team found that this connection was stronger in the oral and skin microbiomes. Interestingly, the skin microbiome was able to pinpoint participants’ ages within an average of 4 years. They also found that relatively few ASVs were required for highly accurate models of each body site. As in previous work, the team found a sex-specific signal in the gut microbiome. However, the researchers did not detect this signal in the mouth or skin microbiome.
The researchers further investigated the skin to determine whether models trained on one body site of the skin could replicate on another. From this, they discovered that models of microbiome age for the forehead could be cross-trained on the palm, and vice versa. This is important as it means that researchers in the future can combine these skin sites when assessing factors leading to microbiome ageing.
Next, the team examined which taxa contributed to the age prediction model. They found that ASVs in the young belonged to more abundant and prevalent taxa than in the elderly. This reflects the change in skin physiology as we age and the associated alterations in microbiota. Therefore, these ASVs could be potential indicators of microbial shifts associated with ageing. This also has implications for microbiome-targeted therapeutic strategies to prevent ageing.
Overall, these results highlight that microbiome studies using machine learning techniques can help accurately indicate ageing. Microbiome studies could help indicate the role of the microbiome in altering the ageing process and also in the susceptibility for age-related diseases. This in turn could lead to microbially-based interventions that could modify the ageing process and impact age-related diseases.
Image credit: By Image Team – canva.com