Utility of a quantitative approach to microbial dysbiosis using machine learning in an African American cohort with self-reported hair loss

Objective Hair loss is a common issue that affects a large proportion of the population, leading to lower self-confidence and quality of life. Microbial dysbiosis of the scalp has been shown to be associated with several different disorders leading to hair loss. Though several “microbiome friendly” cosmetic treatments are currently on the market, there is no agreement on the best technique for assessing dysbiosis leading to a lack of scientific rigor for quantifying the effective treatments. To help address this, the association between self-perceived hair loss and the scalp microbiome in an African-American cohort (n=36) was investigated. Methods Using a self-controlled design, swabs were collected from both “sparse” and “normal” scalp sites. The scalp microbiome was characterized via 16S rRNA gene sequencing and a dysbiosis score was calculated based on the proportion of all taxa within the samples. Further, we identified the taxa that contributed most to abnormal or dysbiotic hair sites using a machine learning random forest classifier and a negative binomial mixed effects model. Results The dysbiosis index is sensitive to participants self-assessment of hair loss and interindividual variation. We found a core set of operational taxonomic units (OTUs) assigned to 7 genera that significantly contributed to increased scalp dysbiosis. Conclusion This study demonstrates that self-perceived hair loss is associated with significant and measurable alterations in the scalp microbiome using, making the reported dysbiosis index a practical tool that may be used to assess microbiome changes following cosmetic or medical interventions for hair loss and other microbiome-associated disorders.