Deep Hair Phenomics: Implications in Endocrinology, Development, and Aging

Area of Science:

• Biomedical Science
• Computer Vision
• Genetics

Background:

• Hair quality is a key health indicator, but current assessment methods are limited in throughput and quantification.
• Assessing individual hair fiber characteristics is challenging due to technical limitations.

Purpose of the Study:

• To develop a high-throughput, quantitative method for analyzing individual hair fibers using deep learning and computer vision.
• To explore the impact of hormonal signaling, genetic modifications, and aging on hair follicle output and hair phenotypes.

Main Methods:

• A novel deep learning-based computer vision approach was created to quantify individual hair fibers.
• The tool can distinguish and extract overlapping fibers for multivariate feature analysis (length, width, color).
• Applied the tool to analyze hair phenotypes in mice across different conditions and life stages.

Main Results:

• Generated high-resolution, single-hair phenomes, revealing diverse hair conditions.
• Identified hair phenotypes associated with hormonal, developmental, and aging-related alterations in mouse fur.
• Demonstrated the tool's efficacy in characterizing factors that modulate hair follicles.

Conclusions:

• The developed deep hair phenomics tool offers a high-throughput solution for quantitative hair analysis.
• This technology enables the characterization of factors influencing hair follicle biology.
• The approach shows potential for novel diagnostic methods by analyzing hair fibers for disease detection.