Deciphering the molecular basis of skin color variation through transcriptomics and machine learning

Skin color is one of the most diverse human traits, but the understanding of the complex genetic mechanisms behind this variation remains incomplete. This study investigated the genetic basis of constitutive skin pigmentation using the Individual Typology Angle (ITA), an objective colorimetric measure. Comparative gene expression between light and dark skin, identified 265 significantly modulated genes. These genes clustered around key pathways, including pigmentation/melanogenesis, antioxidant/stress responses, lipid metabolism (including arachidonic acid and diacylglycerol pathways), and interferon-gamma signaling. As expected, genes involved in pigmentation were upregulated in darker skin. Interestingly, darker skin also exhibited higher expression of antioxidant and detoxification genes like GSTM3 and AKR1B10, suggesting enhanced protection against environmental stressors. Differential expression of genes involved in lipid metabolism has shown potential roles for prostaglandin F2α and diacylglycerol in pigmentation. Furthermore, interferon-gamma signaling, crucial for immune defense and antimicrobial responses, appeared inhibited in darker skin. Finally, a machine learning approach, identified a 25-gene signature for predicting ITA. This signature included known pigmentation genes and novel candidates like GSTM3, PMP22, ENGASE, and SPATS2L. This study provides deeper insights into the intricate interplay of genes and pathways influencing skin pigmentation and its response to environmental factors, laying the groundwork for personalized skincare development.